CN112462704A - Mixed flow batch scheduling optimization method for sensor workshop production - Google Patents

Mixed flow batch scheduling optimization method for sensor workshop production Download PDF

Info

Publication number
CN112462704A
CN112462704A CN202011290641.7A CN202011290641A CN112462704A CN 112462704 A CN112462704 A CN 112462704A CN 202011290641 A CN202011290641 A CN 202011290641A CN 112462704 A CN112462704 A CN 112462704A
Authority
CN
China
Prior art keywords
batch
sensor
machine
ant
path
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.)
Pending
Application number
CN202011290641.7A
Other languages
Chinese (zh)
Inventor
顾文斌
陈泽宇
骆第含
李沛霖
苑明海
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.)
Changzhou Campus of Hohai University
Original Assignee
Changzhou Campus of Hohai University
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 Changzhou Campus of Hohai University filed Critical Changzhou Campus of Hohai University
Priority to CN202011290641.7A priority Critical patent/CN112462704A/en
Publication of CN112462704A publication Critical patent/CN112462704A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00—Programme-control systems
    • G05B19/02—Programme-control systems electric
    • G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
    • G05B19/41865—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by job scheduling, process planning, material flow
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B2219/00—Program-control systems
    • G05B2219/30—Nc systems
    • G05B2219/32—Operator till task planning
    • G05B2219/32252—Scheduling production, machining, job shop
    • 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
    • Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Landscapes

  • Engineering & Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Manufacturing & Machinery (AREA)
  • Quality & Reliability (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • General Factory Administration (AREA)

Abstract

The invention discloses a mixed flow batch scheduling optimization method for sensor workshop production, which solves the problem of batch scheduling optimization scheduling of a mixed flow workshop for a sensor production process by using the minimization of the maximum completion time as an optimization target according to the processing characteristics of the sensor production workshop and the current situation of workshop scheduling. Firstly, a scheduling model of a sensor production workshop is constructed, and secondly, an improved ant colony optimization algorithm is integrated into a batch strategy to solve a scheduling optimal sequence. The mixed flow batch scheduling optimization method for sensor workshop production provided by the invention can play roles in optimizing distribution and improving benefits for resource arrangement, capability balance, quality management, cost and delivery date control of sensor production enterprises, and makes correct technical and management decisions for informatization, standardization and automation construction of enterprises, so that the operation efficiency of manufacturing enterprises is improved and benefits are obtained to the maximum.

Description

Mixed flow batch scheduling optimization method for sensor workshop production
Technical Field
The invention relates to a mixed flow batch scheduling optimization method for sensor workshop production, and belongs to the technical field of communication.
Background
According to the characteristics of the sensor workshop Problem, the Problem is classified into a Batch Scheduling Problem (Batch Scheduling protocol of Hybrid Flow-Shop, BSPHFS) of the Hybrid Flow Shop. The problem is brought forward based on the practical production industrial background of semiconductors, steel, television liquid crystals and the like. Can be generally described as: the workshop needs to produce X types of products or X orders, each type of product or each order is provided with a plurality of workpieces, each workpiece comprises a plurality of continuous production stages with the same constraint, each production stage is provided with at least two or more processing machines, a plurality of workpieces are processed, the processing process routes of the workpieces are the same, at the moment, batches of the workpieces need to be arranged simultaneously, parallel equipment which can be used for processing is arranged in each procedure of the batches after the batches, and the batches which need to be processed on the equipment are sequenced. Therefore, various performance indexes of the production and manufacturing system reach the optimum, the operation efficiency of the manufacturing system is improved, and the production rhythm is accelerated. A batch scheduling hybrid flow plant is shown in figure 1.
Disclosure of Invention
The invention aims to overcome the technical defects in the prior art, solve the technical problems and provide a mixed flow batch scheduling optimization method for sensor workshop production.
The invention specifically adopts the following technical scheme: a mixed flow batch scheduling optimization method for sensor workshop production comprises the following steps:
constructing a dispatching model of a sensor production workshop; and solving the scheduling optimal sequence by adopting a batch strategy and integrating an improved ant colony optimization algorithm.
As a preferred embodiment, the constructing a scheduling model of the sensor production plant includes:
constructing an objective function with the maximum completion time minimized as an objective:
Figure RE-GDA0002908900520000021
wherein X is 1, …, X; s is 1, …, S; n is 1, …, N,
Figure RE-GDA0002908900520000022
represents batch Fx,s,nThe machining end time of (1); equation (1) represents the objective of scheduling optimization to minimize the total completion time.
As a preferred embodiment, the constructing the scheduling model of the sensor production plant further includes:
setting a constraint condition for the objective function:
Figure RE-GDA0002908900520000023
x＝1,…,X；s＝1,…,S；n＝1,…,Nx (2)
in the formula (2), Fx,s,nN-th batch representing the s-th process of the x-th sensor order, where N is 1,2, …, Nx，NxTotal number of lots divided for xth sensor order, JxRepresenting the X-th sensor order, wherein X is 1,2 …, X, wherein X is the total number of sensor orders; formula (2) represents a meaningful process quantity constraint, that is, after each order is batched, the sum of the quantities of the sensors contained in each batch is equal to the quantity of the sensors contained in the order;
Figure RE-GDA0002908900520000024
x＝1,…,X；s＝1,…,S；n＝1,…,Nx
Figure RE-GDA0002908900520000025
in the formula (3), the reaction mixture is,
Figure RE-GDA0002908900520000026
is represented by Fx，s,nProcessing Time of (TU)x,s,mDenotes JxS single sensor processing time on machine, Sum (F)x,s,n) Is represented by Fx,s,nNumber of sensors involved, W (F)x,s,n，Ms,m) Represents a variable of 0 to 1, i.e. 1 represents Fx,s,nBatch process batching machinem, equation (3) indicates that the processing time of a batch of sensors is equal to the processing time of a single sensor multiplied by the number of sensors contained in the batch;
Figure RE-GDA0002908900520000031
x＝1，…,X；s＝1，…,S；n＝1,…,Nx
Figure RE-GDA0002908900520000032
in the formula (4), the reaction mixture is,
Figure RE-GDA0002908900520000033
represents batch Fx,s,nThe end time of the processing of (1),
Figure RE-GDA0002908900520000034
is represented by Fi,s,nThe time required for the processing of (a),
Figure RE-GDA0002908900520000035
is represented by Fx,s,nIf < M indicates a partial order relationship between the sub-batches, if
Figure RE-GDA0002908900520000036
The processing sequence in the sensor shop precedes
Figure RE-GDA0002908900520000037
And no other batches processed, then
Figure RE-GDA0002908900520000038
Note the book
Figure RE-GDA0002908900520000039
Equation (4) represents that the completion time of the lot is equal to the sum of the processing time and the processing start time of the lot;
Figure RE-GDA00029089005200000310
x＝1,…,X；s＝1,…,S；n＝1,…,Nx；
Figure RE-GDA00029089005200000311
in the formula (5), the reaction mixture is,
Figure RE-GDA00029089005200000312
is represented by Fx,s,nThe machining start time of (1) is,
Figure RE-GDA00029089005200000313
represents batch Fx,s,nThe machining end time of (1); the formula (5) represents the maximum time value of the starting processing time of the batch not earlier than the previous batch on the same production line and the finishing time of the previous process of the same batch;
Figure RE-GDA00029089005200000314
x＝1,…,X；s＝1,…,S；n＝1,…,Nx； (6)
in the formula (6), Fx,s,nN-th batch representing the s-th process of the x-th sensor order, where N is 1,2, …, Nx，NxTotal number of lots divided for xth sensor order, Ms,m: m machine processed in the s process in the sensor workshop, wherein M is 1,2, …, MsEquation (6) represents the occupancy constraint of the machine, i.e. all sensors of the same batch can only be processed on the same machine;
Figure RE-GDA0002908900520000041
x＝1,…,X；s＝1,…,S-1；
Figure RE-GDA0002908900520000042
n＝1,…,Nx； (7)
in formula (7), Sum (F)x,s,n) Is represented by Fx,s,nNumber of sensors involved, Jx,s: the S-th process of the x-th sensor order, S ═ all sensor sets of {1,2,3, … …, S }, where S is the total number of processes,
Figure RE-GDA0002908900520000043
a sensor lot indicating a process at a certain time; equation (7) represents a batch constraint relationship, and if the sum of the number of sensors in all completed batches of the s-th process of the x-th sensor order at a certain time is Q, the number of sensors in all completed batches of the s + 1-th process of the x-th sensor order at the certain time should not be greater than Q.
As a preferred embodiment, the step of solving the scheduling optimal sequence by using the improved ant colony optimization algorithm integrated with the batch strategy specifically includes the following steps:
step SS 1: acquiring task batching results and sequencing;
step SS 2: generating a device allocation scheme;
step SS 3: acquiring a processing task on each device, and sequencing processing of each batch on the machine;
step SS 4: determining the optimal path and scheme obtained by searching in the batch sorting layer, ant a2(s) search for a path of
Figure RE-GDA0002908900520000044
And according to the search path length Lmax2Length update scheduling scheme if e2<O2Otherwise, the process goes to step SS2, otherwise, the process goes to step SS 5.
Step SS 5: if the optimal extreme value of 5 sub-optimization is continuously
Figure RE-GDA0002908900520000045
If no change exists, the step SS6 is switched to introduce a catastrophe factor to change the catastrophe factor to a new path, otherwise, the batch sequencing ant colony optimization is finished if the iteration times are finished,go to step SS 7;
step SS 6: introducing a catastrophe factor, changing the concentration of pheromone on a path of a machine distribution layer, and suddenly reducing or reducing the hormone to 0, thereby searching the optimal path in the global range again;
step SS 7: determining an optimal path and a scheme searched by a machine distribution layer algorithm;
step SS 8: if D > D, the algorithm stops; otherwise, the step SS9 is carried out;
step SS 9: and introducing a catastrophe factor, changing the concentration of pheromone on the order batch layer path, and suddenly reducing or reducing the hormone to 0, so as to search the optimal path in the global scope again.
As a preferred embodiment, step SS1 specifically includes: generating an ant a according to the number of the sensors contained in each order1Dividing the order into a set of quanta of batches Wi(i ═ 1,2, …, N), ant a1Randomly selecting a batch W from a batch setiBeginning to traverse, set ant a1The traversed task counter is q1Storing the traversed task as 1
Figure RE-GDA0002908900520000051
Ant a1And then according to the state transition rule, never traversing the task set
Figure RE-GDA0002908900520000052
To select the next task, q1＝q1+ 1; then repeating the steps to complete all the tasks in a traversing way, generating a batch task sequence scheme, and recording the sequence scheme1＝e1+1。
As a preferred embodiment, step SS2 includes: selecting a machine for available processing for each process of each batch, specifically comprising: initializing available capacity K of each process machinem(M is 1,2, …, M), one ant a corresponding to the procedure is generated2(s) and selecting the rule as ant a according to the machine state2(s) initializing a batch node selection machine, setting ant a1Traversed batchesThe counter is q 21, ant a2(s) at the q-th2Step, selectable machine set Mx,s(ii) a From set M according to state transition rulesx,sIf the machine is selected, record q2＝q2+1, updating the capacity of the machines of the process s, continuously searching for the selection of a processing machine for each batch of each process, thus obtaining the final machine allocation scheme, note e2＝e2+1。
As a preferred embodiment, step SS3 specifically includes:
step SS 31: initializing a process s, and acquiring a machine set M of the process ssTo generate an ant a3(m) determining a sequence batch on machine m by a batch transfer rule corresponding to one machine in the set of machines;
firstly, a certain processing batch on a certain machine is randomly selected for each ant as a node for starting traversal: let process s equal to 1, generate an ant a3(m) corresponding to one machine in the machine set to obtain a task set F of all batches on the corresponding machinen(N-1, 2, … N), all ants a3(m) random slave batch task set FnSelecting a batch; let the already-visited batch counter q3Storing the traversed batch set as 1
Figure RE-GDA0002908900520000061
Then the ant a3(m) a set of unsearched batches
Figure RE-GDA0002908900520000062
Then from
Figure RE-GDA0002908900520000063
Select a certain batch of nodes and record q3＝q3+ 1; continuously repeating the process will traverse all the batch nodes, thereby obtaining the batch sequence, and recording e3＝e3+1；
Step SS 32: a decoding process, wherein a batch sequence dispatching result of the machine set is obtained;
(1) setting the initial process as s ═1, initializing path length Qmax3Let Q max30; otherwise, entering (2);
(2) assembling M from machinessExtracting machine M, calculating the machine completion time according to the formulas (2.4) - (2.7), and updating the machine set Ms；
(3) If it is
Figure RE-GDA0002908900520000064
Then Q is updatedmax3If, if
Figure RE-GDA0002908900520000065
Feeding in (4), otherwise, directly feeding in (4);
(4) if it is
Figure RE-GDA0002908900520000066
Step SS33 is entered; otherwise, returning to the step (2);
step SS 33: calculating ant a3(m) searching for the path length L of the device sequencemax3(ii) a To a3(m) locally searching the searched equipment sequence to obtain a neighborhood sequence, comparing the path lengths of the neighborhood sequence and the original sequence, and taking the shortest path L'max3(ii) a Judging whether all the machines of all the processes are scheduled, if s<S, if S is equal to S +1, the process proceeds to step SS 31; otherwise, judging whether each ant has traveled all the nodes of the batch, if e3<O3If yes, the ant still has no journey, and the step is switched to SS 31; otherwise, the step SS34 is carried out;
step SS 34: if the optimal extreme value of 5 sub-optimization is continuously
Figure RE-GDA0002908900520000071
If no change exists, the step SS35 is switched to introduce a catastrophe factor to change the catastrophe factor to a new path, otherwise, the iteration times are finished to finish the batch sorting ant colony algorithm of the equipment, and the step SS4 is switched to;
step SS 35: introducing a catastrophe factor, changing the concentration of pheromone on the path of the batch sequencing layer, and suddenly reducing or reducing the hormone to 0, thereby re-searching the optimal path in the global range;
Figure RE-GDA0002908900520000072
then
Figure RE-GDA0002908900520000073
And updating the current optimal scheme; the pheromone updating method comprises the following steps: attenuation of pheromone on long path obtained by search to tauij(t+n)＝(1-ρ)·τij(t)+ΔτijConversely, the shortest time-to-completion pathway increases pheromones to
Figure RE-GDA0002908900520000074
Wherein a is3(min) the shortest ant length is selected,
Figure RE-GDA0002908900520000075
proceed to step SS 31.
As a preferred embodiment, step SS6 specifically includes: if it is
Figure RE-GDA0002908900520000076
Then
Figure RE-GDA0002908900520000077
And updating the machine allocation plan; the pheromone updating method comprises the following steps: attenuation of pheromone on long path obtained by search to taukm(t+n)＝(1-ρ)·τkm(t)+Δτkm(ii) a Conversely, the shortest completion time pathway increases its pheromone to
Figure RE-GDA0002908900520000078
Wherein a is3(min) the shortest ant length is selected,
Figure RE-GDA0002908900520000079
proceed to step SS 2.
As a preferred embodiment, step SS7 specifically includes: ant a1The searched path is
Figure RE-GDA00029089005200000710
And according to the search path length Lmax1A length update scheduling scheme; to a1The path of the sequence is locally searched to obtain a neighborhood sequence, the path lengths of the neighborhood sequence and the original sequence are compared, and the shortest path L 'is taken'max1. If e1<O1Then, go to step SS 1; otherwise, the process goes to step SS 8.
As a preferred embodiment, step SS9 specifically includes: if it is
Figure RE-GDA00029089005200000711
Then
Figure RE-GDA00029089005200000712
And updating the batch schedule; the pheromone updating method comprises the following steps: attenuation of pheromone on long path obtained by search to tauni(t+n)＝(1-ρ)·τni(t)+Δτkm(ii) a Conversely, the shortest completion time pathway increases its pheromone to
Figure RE-GDA00029089005200000713
Wherein a is1(min) the shortest ant length is selected,
Figure RE-GDA0002908900520000081
proceed to step SS 1.
The invention achieves the following beneficial effects: the mixed flow batch scheduling optimization method for sensor workshop production mainly aims at the batch scheduling problem of the mixed flow workshop, establishes a static mathematical model and a batch strategy for batch scheduling of the mixed flow workshop, designs a three-layer structure alternative iterative catastrophe ant colony algorithm, and performs optimization solution by using a multi-order ant colony algorithm to perform batch-batch assignment machining on a product (order) and sequencing the machining batches on corresponding machines, wherein the algorithm takes the maximum completion minimum time as a target. And then, evaluating the effectiveness and superiority of the algorithm respectively through example simulation, and showing that the algorithm is relatively superior in processing capacity, optimization or time through results. And finally, combining with an actual calculation example to obtain an optimal scheduling scheme.
Drawings
FIG. 1 is a schematic view of a mixing plant of the present invention;
FIG. 2 is an exemplary diagram of a sensor shop order (product) batch strategy of the present invention;
FIG. 3 is a schematic diagram of a sensor shop scheduling process of the present invention;
FIG. 4 is a Gantt chart of a catastrophe ant colony algorithm with three-layer structure alternate iteration;
FIG. 5 is an algorithmic flow diagram of a preferred embodiment of the present invention;
FIG. 6 is a graph of the effect of different population sizes on algorithm performance of the present invention;
FIG. 7 is a graph of the effect of different process numbers on algorithm performance of the present invention;
FIG. 8 is a graph of the effect of different machine counts of the present invention on algorithm performance;
FIG. 9 is a comparison of ARPD values for different algorithms for the same example;
fig. 10 is a gantt chart of scheduling.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
Example 1: according to the characteristics of the sensor workshop Problem, the Problem is classified into a Batch Scheduling Problem (Batch Scheduling protocol of Hybrid Flow-Shop, BSPHFS) of the Hybrid Flow Shop. The problem is brought forward based on the practical production industrial background of semiconductors, steel, television liquid crystals and the like. Can be generally described as: the workshop needs to produce X types of products or X orders, each type of product or each order is provided with a plurality of workpieces, each workpiece comprises a plurality of continuous production stages with the same constraint, each production stage is provided with at least two or more processing machines, a plurality of workpieces are processed, the processing process routes of the workpieces are the same, at the moment, batches of the workpieces need to be arranged simultaneously, parallel equipment which can be used for processing is arranged in each procedure of the batches after the batches, and the batches which need to be processed on the equipment are sequenced. Therefore, various performance indexes of the production and manufacturing system reach the optimum, the operation efficiency of the manufacturing system is improved, and the production rhythm is accelerated. A batch scheduling hybrid flow plant is shown in figure 1 below.
1.2 problem hypothesis: according to the condition of a workshop of a sensor manufacturing system, various factors influencing the production operation of the workshop in the process of researching the problems can be easily found, so that the solution of many practical problems is very complicated. In order to solve the actual problem, the actual problem is usually theoretically simplified, interference of each influence factor is analyzed, and a rationalization constraint condition is constructed, and the process is generally called as hypothesis on the problem, namely, constraint conforming to the actual meaning of a research model is made, so that the actual problem becomes investigatable, and key influence factors are not distorted.
Therefore, for the sensor shop scheduling problem, the following assumptions are made:
selecting machining rights of all workpieces at the starting moment;
the process processing time is determined by the capacity of the equipment;
the batch changing time of different batches of processing is considered in the conveying time of the workpieces;
fourthly, after the machine starts to process the batch of workpieces, the machining process is not allowed to stop or be interrupted;
the machine has to finish processing the workpieces of the current batch, and can process the workpieces of the next batch;
sixthly, the working procedures of workpieces in different batches are not successively restricted;
the same batch of workpieces is not allowed to appear on different machines for processing;
the machining processes of workpieces in the same batch are sequentially restricted;
the self-lifting ensures the sufficient material;
with the sum of the batches of each sublot equal to the sum of the number of order sensors.
The main problems of the sensor production workshop are that the processing period is long, and the actual processing time deviates from the completion time of a production scheduling plan, so that the subsequent production plan is unsmooth. For this purpose, it is proposed herein to minimize the maximum completion time.
An objective function:
Figure RE-GDA0002908900520000101
wherein X is 1, …, X; s is 1, …, S; n is 1, …, N,
Figure RE-GDA0002908900520000102
represents batch Fx,s,nThe machining end time of (1). Equation (1) represents the objective of scheduling optimization to minimize the total completion time.
Constraint conditions are as follows:
Figure RE-GDA0002908900520000103
x＝1,…,X；s＝1,…,S；n＝1,…,Nx (2)
in the formula (2), Fx，s，nN-th lot (N is 1,2, …, N) representing the s-th process of the x-th sensor orderx),NxTotal number of lots divided for xth sensor order, JxThe X-th sensor order (X ═ 1,2 …, X) is indicated, where X is the total number of sensor orders. Equation (2) represents a meaningful process quantity constraint, i.e., after each order is batched, the sum of the quantities of sensors contained in each batch is equal to the quantity of sensors contained in the order.
Figure RE-GDA0002908900520000104
x＝1,…,X；s＝1,…,S；n＝1,…,Nx
Figure RE-GDA0002908900520000105
In the formula (3), the reaction mixture is,
Figure RE-GDA0002908900520000111
is represented by Fx,s,nProcessing Time of (TU)x,s,mDenotes JxS single sensor processing time on machine, Sum (F)x,s,n) Is represented by Fx,s,nNumber of sensors involved, W (F)x，s，n，Ms，m) Represents a variable of 0 to 1, i.e. 1 represents Fx，s，nBatch process batches are processed on machine m, equation (3) meaning that the processing time of a batch of sensors is equal to the processing time of an individual sensor multiplied by the number of sensors contained in the batch.
Figure RE-GDA0002908900520000112
x＝1，…,X；s＝1,…,S；n＝1,…,Nx
Figure RE-GDA0002908900520000113
In the formula (4), the reaction mixture is,
Figure RE-GDA0002908900520000114
represents batch Fx,s,nThe end time of the processing of (1),
Figure RE-GDA0002908900520000115
is represented by Fi,s,nThe time required for the processing of (a),
Figure RE-GDA0002908900520000116
is represented by Fx,s,nIf < M indicates a partial order relationship between the sub-batches, if
Figure RE-GDA0002908900520000117
The processing sequence in the sensor shop precedes
Figure RE-GDA0002908900520000118
And no other batches processed, then
Figure RE-GDA0002908900520000119
For convenient calculation, note
Figure RE-GDA00029089005200001110
Equation (4) represents that the completion time of the lot is equal to the sum of the processing time and the processing start time of the lot.
Figure RE-GDA00029089005200001111
x＝1,…,X；s＝1,…,S；n＝1,…,Nx；
Figure RE-GDA00029089005200001112
In the formula (5), the reaction mixture is,
Figure RE-GDA00029089005200001113
is represented by Fx,s,nThe machining start time of (1) is,
Figure RE-GDA00029089005200001114
represents batch Fx,s,nThe machining end time of (1); equation (5) represents the maximum time value of the starting time of a batch not earlier than the previous batch on the same production line and the finishing time of the previous process of the same batch.
Figure RE-GDA00029089005200001115
x＝1,…,X；s＝1,…,S；n＝1,…,Nx； (6)
In the formula (6), Fx,s,nN-th lot (n is 1,2, …, Nx) indicating the s-th process of the x-th sensor order, Nx being the total number of lots into which the x-th sensor order is divided, Ms, m: m-th machine (M1, 2, …, M) processed in the s-th process in the sensor shops) Equation (6) represents the occupancy constraint of the machine, i.e. all sensors of the same batch can only be processed on the same machine.
Figure RE-GDA0002908900520000121
x＝1,…,X；s＝1,…,S-1；
Figure RE-GDA0002908900520000122
n＝1,…,Nx； (7)
In formula (7), Sum (F)x,s,n) Is represented by Fx,s,nNumber of sensors involved, Jx,s: all sensor sets of the S-th process (S ═ {1,2,3, … …, S }) of the x-th sensor order, where S is the total number of processes,
Figure RE-GDA0002908900520000123
a sensor lot indicating a process at a certain time; equation (7) represents a batch constraint relationship, and if the sum of the number of sensors in all completed batches of the s-th process of the x-th sensor order at a certain time is Q, the number of sensors in all completed batches of the s + 1-th process of the x-th sensor order at the certain time should not be greater than Q.
1.4 scheduled batch strategy: according to the characteristics of a sensor production workshop, the quantity required by production is very large, so that the production capacity requirement of single workpieces cannot be met in a production mode, a batch thought is provided, namely, a batch of large batches of workpieces is formed into n batches in a certain quantity, each batch has certain batch, production time and completion time, and finally, the workpieces are processed according to the batches, so that the quantity of the workpieces processed in unit time is increased, the production rhythm is accelerated, and the production capacity is improved. But the batching requirements are different due to different actual workshop environments, and the batching aims to generate batch batches which can adapt to the workshop environment, so that the limited workshop resources are reasonably distributed to the maximum extent.
In practice, the production period is influenced by the number of batches, and the batches after the batches are in close relation to the actual production operation efficiency of a workshop. Research shows that the batch size has influence on the energy consumption and the production period of the production system, and in the batch production process, when the number of sub-batch workpieces is too large, the current equipment is in high-energy operation processing, and the subsequent equipment is in a waiting state for a long time, so that the production period is increased, and the energy consumption of some equipment is too large; when the number of the sub-batches of workpieces is too small, the number of the batches is increased, so that the problem search range is enlarged, the calculation performance of the algorithm is influenced, the batch changing time is increased, and the equipment processing times are frequent, so that the operation energy consumption is increased. Therefore, the proper amount of batching can effectively shorten the production period, save the energy consumption of equipment and simultaneously improve the operation efficiency of a workshop.
This section presents a flexible batching strategy for mass produced sensor plants. During batching, workpieces in the same order are divided into a plurality of sub-batches according to a certain quantity, and then all the sub-batches are arranged in a processing sequence. Taking a certain 3 orders in a sensor plant as an example, each order contains 16 products. FIG. 2 is a diagram showing an example of a sensor shop order batch strategy, with order batch first: there are 3 orders A, B, C, each of which is divided into 4 tasks and marked as A, with 4 workpieces as a sub-batch task1，A2，A3,A4And the other tasks are analogized, the batch processing sequencing is carried out after the batch is finished, and a task sequencing sequence is randomly generated and is B1B2A1C1A1B3C2C3A3B4C4A4The number of batches was 12. Experiments were also performed on the batch quantities of the actual sensor production plant to find quantities adapted to the plant capacity.
2-three-layer structure alternative processing catastrophe ant colony algorithm solution mixed flow batch scheduling model
2.1 static mixed flow batch scheduling three-layer structure alternate iteration catastrophe ant colony algorithm
Aiming at the batch problem of the sensor mixing flow shop, a three-layer structure alternative iteration catastrophe ant colony algorithm is provided. The ant colony algorithm has remarkable effects on solving the optimized combination, finding the optimal path, sequencing and the like. However, as the scale of the research problem is larger and larger, the range of the problem solving is wider and wider, the ant colony algorithm has the disadvantages that the early-maturing phenomenon occurs in rapid convergence, and the problem is solved by considering, so that the catastrophe factor is introduced. In the biological evolution process, scientists find that catastrophe is an important influencing factor in the biological evolution process. The growth and evolution speed of everything is constant, from rapid evolution to evolution stop, and then the change of environment and life habits caused by catastrophe enters a new growth period. It is then conceivable that it is also possible for ants to find food through catastrophe so that ants start to find new paths to find food. Therefore, a catastrophe factor is introduced into the ant colony algorithm, namely the catastrophe factor is used for changing the quantity of the pheromone in the path to reduce the quantity to zero, so that the algorithm jumps out of the current path and starts to search a new path again, and then iteration is carried out continuously to find the optimal path in the full range. The ant colony algorithm with the introduced catastrophe factors can avoid the premature phenomenon, so that the algorithm becomes more optimized. Considering the complexity and diversity of the batch scheduling problem, a three-layer structure alternative iteration catastrophe ant colony algorithm is designed.
According to the production characteristics of the sensor workshop, the method mainly completes three parts of work for dispatching and optimizing the workshop in detail: firstly, batching orders according to a certain batch; distributing processing equipment for each process of each processing task, and sequencing the batches. FIG. 3 is a diagram of a sensor shop scheduling process.
The scheduling process adopts a three-layer structure alternately nested catastrophe ant colony algorithm. The upper layer is used for establishing an order optimization batch scheme (a workpiece preparation scheme) by using a catastrophe ant colony algorithm, the middle layer and the lower layer are used for optimizing and dispatching the sub-batches (the arrangement of processing equipment of the workpieces and the sequencing of processing sequences on the equipment), the batch scheme of the upper layer is an object of the middle-lower layer dispatching, the middle-lower layer is used for searching and distributing the equipment by using the catastrophe ant colony optimization algorithm, then the batches on each machine are searched for the optimal sequencing, and the batch scheme of the upper layer is the implementation of a batch strategy. Fig. 3 is a schematic diagram of a sensor shop scheduling process, an upper layer in fig. 3 is an order batch layer, each square represents a task batch, batches are equally batched through a batch strategy, and ants sequentially traverse nodes to obtain a sequence of batch tasks, so that several selectable processing batch schemes are obtained. According to the batch scheme of the upper layer, the ants in the middle layer search all batches, and equipment which can be used for processing is arranged for each process of each batch. Each white square in the middle level represents a batch of selectable equipment and the black square represents a batch of selected process equipment, such that equipment allocation plans for all processes in all batches are searched. And the batch sorting in the lower layer obtains all equipment distribution schemes through ant colony algorithm searching, calculates the length of each possible path passed by ants for all the batches on each equipment, and finally finds out the optimal path. In fig. 3, a lower node represents a processable batch on the equipment, ants sequentially traverse all the nodes to obtain a batch ordering scheme of the equipment and calculate the path length of the batch ordering scheme, and finally, an ant colony algorithm is used for evaluating the searched path to obtain an optimal scheduling scheme.
In order to facilitate understanding of the scheduling process, a scheduling process of an optimal scheduling scheme is described by taking an example that a plurality of selectable processing devices are arranged in a process corresponding to 3 required processing processes of 3 orders (products), the orders are marked as { a, B, C }, and the devices are marked as { M1,M2,M3,M4,M5,M6,M7Assuming that each order has 4 tasks, and a batch strategy is adopted to take 2 workpieces as one batch, all the batch results are { A }1,A2,B1,B2,C1,C2In which "A" is1"indicates batch 1 of order (product) A, and so on. The batch sequence obtained by searching the upper-layer catastrophe ant colony algorithm is { A2,B1,C2,C1,A1,B2As an initial object of the device allocation layer. Searching { A } by catastrophe ant colony algorithm in middle layer2,B1,C2,C1,A1,B2Arranging the processing equipment for the 3 procedures of the corresponding batch as M2,M1,M1M2,M1,M2；M3,M4,M4M3,M4,M3；M5,M6,M7,M6,M5,M7The division of the division into working procedures 1,2 and I is carried outSequence 3), the processing lot of the corresponding device can be deduced as { M }1:A1,B1,C2；M2:A2,B2,C1；M3:A2,B2,C1； M4:A1,B1,C2；M5:A1,A2；M6:B1,C1；M7:B2,C2And then searching the lower layer by a catastrophe ant colony algorithm to obtain a batch sequence { B1,C2,A1；A2,C1,B2；A2,C1,B2；B1,C2,A1；A2,A1；B1,C1；C2,B2Push out the processing lot arrangement of the corresponding machine as { M }1:B1,C2,A1；M2:A2,C1,B2；M3:A2,C2,A1；M4:B1,C1,B2；M5:A2,A1；M6:B1,C1(ii) a M, and thus a gantt chart corresponding to the scheduling scheme is shown in fig. 4.
1) Algorithmic process
Sign setting of basic parameters of the catastrophe ant colony algorithm: number of ants Oi(i ═ 1,2,3), pheromone attenuation coefficient δi(i ═ 1,2,3), the number of iterations of the algorithm K, the ant counter ei(i ═ 1,2,3), loop counter t, heuristic factor α. The algorithm flow chart is shown in fig. 5.
Step 1: and acquiring the batch results and sequencing of the tasks.
Generating an ant a according to the number of the sensors contained in each order1Dividing the order into a set of quanta of batches Wi(i ═ 1,2, …, N), ant a1Randomly selecting a batch W from a batch setiBeginning to traverse, set ant a1The traversed task counter is q1Storing the traversed task as 1
Figure RE-GDA0002908900520000161
Ant a1And then according to the state transition rule, never traversing the task set
Figure RE-GDA0002908900520000162
To select the next task, q1＝q1+ 1; then repeating the steps to complete all the tasks in a traversing way, generating a batch task sequence scheme, and recording the sequence scheme1＝e1+1。
Step 2: and generating a device allocation scheme.
Step 2.1 selects the available processing machines for each pass of the batch.
Initializing available capacity K of each process machinem(M is 1,2, …, M), one ant a corresponding to the procedure is generated2(s) and selecting the rule as ant a according to the machine state2(s) initializing a batch node selection machine, setting ant a1The traversed batch counter is q 21, ant a2(s) at the q-th2Step, selectable machine set Mx,s. From set M according to state transition rulesx,sIf the machine is selected, record q2＝q2+1, update process s machine capability. The search is repeated to select a processing machine for each batch of each process. Resulting in a final machine allocation scheme. Note e2＝e2+1。
And step 3: the processing tasks on each equipment are acquired, and then the processing of each batch on the machine is sequenced.
And (4) acquiring a machine distribution scheme according to the step 2, and determining that the individual machine needs to process the batch task. The process of sequencing the batches on each machine is as follows.
Step 3.1: initializing a process s, and acquiring a machine set M of the process ssTo generate an ant a3(m) determining the sequence lot on machine m by the lot transfer rule corresponding to one machine in the set of machines.
Firstly, a certain processing batch on a certain machine is randomly selected for each ant as a node for starting traversal: let process s equal to 1, generate one anta3(m) corresponding to one machine in the machine set to obtain a task set F of all batches on the corresponding machinen(N-1, 2, … N), all ants a3(m) random slave batch task set FnOne batch is selected. Let the already-visited batch counter q3Storing the traversed batch set as 1
Figure RE-GDA0002908900520000171
Then the ant a3(m) a set of unsearched batches
Figure RE-GDA0002908900520000172
Then from
Figure RE-GDA0002908900520000173
Select a certain batch of nodes and record q3＝q3+1. Repeating the process continuously will complete all the batch nodes, thereby obtaining the batch sequence. And record e3＝e3+1。
Step 3.2: and in the decoding process, the batch sequence scheduling result of the machine set is obtained.
(1) If the initial process is s equal to 1, the path length Q is initializedmax3 Let Q max30; otherwise, enter (2).
(2) Assembling M from machinessExtracting machine M, calculating the machine completion time according to the formulas (2.4) - (2.7), and updating the machine set Ms。
(3) If it is
Figure RE-GDA0002908900520000181
Then Q is updatedmax3If, if
Figure RE-GDA0002908900520000182
Feeding in (4), otherwise, directly feeding in (4).
(4) If it is
Figure RE-GDA0002908900520000183
Step 3.3 is entered; otherwise, return to (2).
Step 3.3: counting deviceAnt a3(m) searching for the path length L of the device sequencemax3. To a3(m) locally searching the searched equipment sequence to obtain a neighborhood sequence, comparing the path lengths of the neighborhood sequence and the original sequence, and taking the shortest path L'max3. Judging whether all the machines of all the processes are scheduled, if s<S, if S is equal to S +1, go to step 3; otherwise, judging whether each ant has traveled all the nodes of the batch, if e3<O3If yes, the ant does not have a tour, and the step 3 is carried out; otherwise, go to step 3.4.
Step 3.4: if the optimal extreme value of 5 sub-optimization is continuously
Figure RE-GDA0002908900520000184
And (4) if no change exists, turning to the step 3.5 to introduce a catastrophe factor to change the catastrophe factor to a new path, otherwise, finishing the batch sorting ant colony algorithm of the equipment after the iteration times, and turning to the step 4.
Step 3.5: and introducing a catastrophe factor, changing the concentration of pheromone on the path of the batch sequencing layer, and suddenly reducing or reducing the hormone to 0, so as to search the optimal path in the global range again.
Figure RE-GDA0002908900520000185
Then
Figure RE-GDA0002908900520000186
And updates the current best solution. The pheromone updating method comprises the following steps: attenuation of pheromone on long path obtained by search to tauij(t+n)＝(1-ρ)·τij(t)+ΔτijConversely, the shortest time-to-completion pathway increases pheromones to
Figure RE-GDA0002908900520000187
Wherein a is3(min) the shortest ant length is selected,
Figure RE-GDA0002908900520000188
and (5) turning to the step 3.
And 4, step 4: and determining a batch sorting layer to search to obtain an optimal path and scheme.
Ant a2(s) search for a path of
Figure RE-GDA0002908900520000189
And according to the search path length Lmax2The length update scheduling scheme. If e2<O2And if not, the step 2 is carried out, and otherwise, the step 5 is carried out.
And 5: if the optimal extreme value of 5 sub-optimization is continuously
Figure RE-GDA00029089005200001810
And (4) if no change exists, turning to the step 6 to introduce a catastrophe factor, changing the catastrophe factor to a new path, otherwise, finishing the batch sequencing ant colony optimization after the iteration times are finished, and turning to the step 7.
Step 6: and introducing a catastrophe factor, changing the concentration of pheromone on the path of the distribution layer of the machine, and suddenly reducing or reducing the hormone to 0, so as to search the optimal path in the global scope again.
If it is
Figure RE-GDA0002908900520000191
Then
Figure RE-GDA0002908900520000192
And updates the machine allocation scheme. The pheromone updating method comprises the following steps: attenuation of pheromone on long path obtained by search to taukm(t+n)＝(1-ρ)·τkm(t)+Δτkm(ii) a Conversely, the shortest completion time pathway increases its pheromone to
Figure RE-GDA0002908900520000193
Figure RE-GDA0002908900520000194
Wherein a is3(min) the shortest ant length is selected,
Figure RE-GDA0002908900520000195
and (5) transferring to the step 2.
And 7: and determining the optimal path and scheme searched by the machine distribution layer algorithm.
Ant a1The searched path is
Figure RE-GDA0002908900520000196
And according to the search path length Lmax1The length update scheduling scheme. To a1The path of the sequence is locally searched to obtain a neighborhood sequence, the path lengths of the neighborhood sequence and the original sequence are compared, and the shortest path L 'is taken'max1. If e1<O1Then, turning to the step 1; otherwise, go to step 8.
And 8: if D > D, the algorithm stops; otherwise, go to step 9.
And step 9: and introducing a catastrophe factor, changing the concentration of pheromone on the order batch layer path, and suddenly reducing or reducing the hormone to 0, so as to search the optimal path in the global scope again.
If it is
Figure RE-GDA0002908900520000197
Then
Figure RE-GDA0002908900520000198
And the batch protocol is updated. The pheromone updating method comprises the following steps: attenuation of pheromone on long path obtained by search to tauni(t+n)＝(1-ρ)·τni(t)+ Δτkm. Conversely, the shortest completion time pathway increases its pheromone to
Figure RE-GDA0002908900520000199
Figure RE-GDA00029089005200001910
Wherein a is1(min) the shortest ant length is selected,
Figure RE-GDA00029089005200001911
and (5) transferring to the step 1.
2) Ant state transition rules
And on the upper layer, ants carry out state transfer by adopting a pseudo-random proportion rule, and select a task to be processed next. Probability of state transition:
Figure RE-GDA00029089005200001912
in the formula:
Figure RE-GDA0002908900520000201
is the pheromone level between tasks (n, i);
Figure RE-GDA0002908900520000202
middle layer of each ant a2The state transition probability of(s) is:
Figure RE-GDA0002908900520000203
in the formula:
Figure RE-GDA0002908900520000204
pheromone level of the selected machine for lot (k, m);
Figure RE-GDA0002908900520000205
Fx,s,nrepresenting a batch node n.
Lower layer of each ant a3The state transition probability of (m) is:
Figure RE-GDA0002908900520000206
2.2 Algorithm validation
The three-layer structure alternative iteration catastrophe ant colony algorithm takes Matlab as a simulation environment, and adopts a computer with Win 10: the processor Intel core i7 has a main frequency of 4.0GHz and a memory of 8G. The verification includes two aspects: (1) performance evaluation and comparison; (2) solving the batch scheduling model of the HFS.
Performance evaluation and comparison
The method aims at evaluating the performance of the three-layer structure alternative iterative catastrophe ant colony algorithm, and mainly evaluates the effectiveness and superiority of the three-layer structure alternative iterative catastrophe ant colony algorithm. The effectiveness mainly considers the influence of the parameters of the algorithm and the problem solving parameters on the performance of the algorithm; the advantage aspect is that the optimal result is solved for the same example by comparing with other algorithms.
I. Effect of different test cases on Algorithm Performance
The effectiveness of the algorithm is verified through different parameter settings, and the influence of different population sizes (example 1), different process numbers (example 2) and the number of devices per process (example 3) on the performance of the MACO-TLSAI (Multi Ant organic algorithm of Three-Layer Structure Alternating evaluation) algorithm is mainly considered. The relevant parameters for the experimental test examples are shown in table 1.
TABLE 1 design of different test case parameters
Figure RE-GDA0002908900520000211
Three corresponding results are obtained by testing the three groups of examples, and as can be seen from fig. 6, 7 and 8, the MACO-TLSAI algorithm can quickly converge and approach to an optimal solution along with the number of iterations as the number of ant populations continuously increases; the number of the procedures is increased under other conditions, and the capability of the MACO-TLSAI algorithm for solving the optimal solution of the target is obviously reduced; and thirdly, the population and the working procedures are ensured to be fixed, and the number of selectable processing machines is increased for each working procedure, so that the quality of the target optimal solution can be improved.
Comparative experimental study
Aiming at the superiority evaluation of the catastrophe ant colony algorithm of the three-layer structure alternate iteration, the performance index and the calculation time are mainly used for comparing the ant colony algorithm (MACO-TLSAI) of the three-layer structure alternate iteration with the genetic and simulated annealing algorithm, and the simulation calculation is carried out on the classical arithmetic. The calculation example is 18 calculation examples of different processing time of different processes with different numbers of workpieces; the calculation examples are shown by the number of workpieces and the number of processes, for example, 6X 2 represents 6 workpieces, 2 processes, and T1And T2Examples of different processing times are shown, each example being run 15 times. The comparison results are shown in tables 2 and 3Shown in the figure.
TABLE 2 Algorithm comparative analysis experiment result I
Figure RE-GDA0002908900520000221
Figure RE-GDA0002908900520000231
TABLE 3 Algorithm comparative analysis experiment result two
Figure RE-GDA0002908900520000232
Figure RE-GDA0002908900520000241
According to the analysis of the experimental results in the table, the MACO-TLSAI algorithm is excellent in both the maximum completion time performance index and the average value; for the calculation time, the MACO-TLSAI convergence speed is faster, the calculation time is reduced, and the advantage of the calculation time is more and more obvious along with the increase of the problem scale. Of course, to more clearly see the performance comparison between the different algorithms, an Average Relative Performance Development (ARPD) is defined as equation (12).
Figure RE-GDA0002908900520000242
Figure RE-GDA0002908900520000243
Mean optimum for each run, min: the algorithm obtains the optimal solution in the experiment. The smaller ARPD is, the closer the average value is to the known optimal solution, and the better the algorithm solution result is. The ARPD values of 18 sets of algorithms are given in fig. 9, and it can be seen that the MACO-TLSAI solution result is better than the other two algorithms.
Solving batch scheduling model of mixed flow shop
The MACO-TLSAI algorithm is adopted to solve the batch scheduling problem of the mixed flow water plants, and the scheduling objective is to minimize the maximum completion time. The test example is that 6 products need to be processed in the production of a sensor workshop of a certain company, 8 products need to be processed for each product, the processing process route of each workpiece is the same, 4 procedures of bridge combination 1, welding 1, bridge combination 2 and mechanical testing are required, 5 pieces of bridge combination equipment, 2 pieces of welding machines and 2 pieces of mechanical testing are available, the processing capacity of each piece of equipment is different, and specific data are shown in table 3.
TABLE 3 actual measurement example
Figure RE-GDA0002908900520000251
In the Gantt chart of figure 10,
Figure RE-GDA0002908900520000252
the numbers representing the top represent the order (product) category 6 and the numbers representing the number of batches batched for such order (product), i.e. batch 4, represented as order (product) 6; the different colors in the figure represent different categories. The algorithm divides all orders (products) into 24 batches, resulting in a production cycle 1008.
3 mixed flow batch scheduling optimization method for sensor workshop production
The mixed flow batch scheduling optimization method for sensor workshop production mainly aims at the mixed flow workshop batch scheduling problem, establishes a mixed flow workshop batch scheduling static mathematical model and a batch strategy, designs a three-layer structure alternative iterative catastrophe ant colony algorithm, and performs batch-batch assignment machining on a product (order) by utilizing a multi-order ant colony algorithm with the maximum completion minimum time as a target so as to perform optimization solution on the sequencing of machining batches on corresponding machines. And then, evaluating the effectiveness and superiority of the algorithm respectively through example simulation, and showing that the algorithm is relatively superior in processing capacity, optimization or time through results. And finally, combining with an actual calculation example to obtain an optimal scheduling scheme.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.

Claims (10)

1. A mixed flow batch scheduling optimization method for sensor workshop production is characterized by comprising the following steps:
constructing a dispatching model of a sensor production workshop; and solving the scheduling optimal sequence by adopting a batch strategy and integrating an improved ant colony optimization algorithm.
2. The mixed-flow batch scheduling optimization method for sensor shop production according to claim 1, wherein the constructing a scheduling model of a sensor shop comprises:
constructing an objective function with the maximum completion time minimized as an objective:
Figure FDA0002783691790000011
wherein X is 1, …, X; s is 1, …, S; n is 1, …, N,
Figure FDA0002783691790000012
represents batch Fx,s,nThe machining end time of (1); equation (1) represents the objective of scheduling optimization to minimize the total completion time.
3. The mixed-flow batch scheduling optimization method for sensor shop production according to claim 1, wherein the constructing the scheduling model of the sensor shop further comprises:
setting a constraint condition for the objective function:
Figure FDA0002783691790000013
x＝1,…,X；s＝1,…,S；n＝1,…,Nx (2)
in the formula (2), Fx,s,nN-th batch representing the s-th process of the x-th sensor order, where N is 1,2, …, Nx，NxTotal number of lots divided for xth sensor order, JxRepresenting the X-th sensor order, wherein X is 1,2 …, X, wherein X is the total number of sensor orders; formula (2) represents a meaningful process quantity constraint, that is, after each order is batched, the sum of the quantities of the sensors contained in each batch is equal to the quantity of the sensors contained in the order;
Figure FDA0002783691790000021
x＝1,…,X；s＝1,…,S；n＝1,…,Nx
Figure FDA0002783691790000022
in the formula (3), the reaction mixture is,
Figure FDA0002783691790000023
is represented by Fx,s,nProcessing Time of (TU)x,s,mDenotes JxS single sensor processing time on machine, Sum (F)x,s,n) Is represented by Fx,s,nNumber of sensors involved, W (F)x,s,n，Ms,m) Represents a variable of 0 to 1, i.e. 1 represents Fx,s,nBatch process batches are processed on machine m, equation (3) indicates that the processing time of a batch of sensors is equal to the processing time of a single sensor multiplied by the number of sensors contained in the batch;
Figure FDA0002783691790000024
x＝1,…,X；s＝1,…,S；n＝1,…,Nx
Figure FDA0002783691790000025
in the formula (4), the reaction mixture is,
Figure FDA0002783691790000026
represents batch Fx,s,nThe end time of the processing of (1),
Figure FDA0002783691790000027
is represented by Fi,s,nThe time required for the processing of (a),
Figure FDA0002783691790000028
is represented by Fx,s,nThe machining start time of (1) is,
Figure FDA00027836917900000213
representing partial order relationship between the sub-batches if
Figure FDA0002783691790000029
The processing sequence in the sensor shop precedes
Figure FDA00027836917900000210
And no other batches processed, then
Figure FDA00027836917900000211
Note the book
Figure FDA00027836917900000212
Equation (4) represents that the completion time of the lot is equal to the sum of the processing time and the processing start time of the lot;
Figure FDA0002783691790000031
x＝1,…,X；s＝1,…,S；n＝1,…,Nx；
Figure FDA0002783691790000032
in the formula (5), the reaction mixture is,
Figure FDA0002783691790000033
is represented by Fx,s,nThe machining start time of (1) is,
Figure FDA0002783691790000034
represents batch Fx,s,nThe machining end time of (1); the formula (5) represents the maximum time value of the starting processing time of the batch not earlier than the previous batch on the same production line and the finishing time of the previous process of the same batch;
Figure FDA0002783691790000035
x＝1,…,X；s＝1,…,S；n＝1,…,Nx； (6)
in the formula (6), Fx,s,nN-th batch representing the s-th process of the x-th sensor order, where N is 1,2, …, Nx，NxTotal number of lots divided for xth sensor order, Ms,m: m machine processed in the s process in the sensor workshop, wherein M is 1,2, …, MsEquation (6) represents the occupancy constraint of the machine, i.e. all sensors of the same batch can only be processed on the same machine;
Figure FDA0002783691790000036
x＝1,…,X；s＝1,…,S-1；
Figure FDA0002783691790000037
n＝1,…,Nx； (7)
in formula (7), Sum (F)x,s,n) Is represented by Fx,s,nNumber of sensors involved, Jx,s: the S-th process of the x-th sensor order, S ═ all sensor sets of {1,2,3, … …, S }, where S is the total number of processes,
Figure FDA0002783691790000041
a sensor lot indicating a process at a certain time; equation (7) represents a batch constraint relationship, and if the sum of the number of sensors in all completed batches of the s-th process of the x-th sensor order at a certain time is Q, the number of sensors in all completed batches of the s + 1-th process of the x-th sensor order at the certain time should not be greater than Q.
4. The mixed-flow batch scheduling optimization method for sensor workshop production according to claim 1, wherein the solving of the scheduling optimal sequence by adopting a batch strategy and incorporating an improved ant colony optimization algorithm specifically comprises the following steps:
step SS 1: acquiring task batching results and sequencing;
step SS 2: generating a device allocation scheme;
step SS 3: acquiring a processing task on each device, and sequencing processing of each batch on the machine;
step SS 4: determining the optimal path and scheme obtained by searching in the batch sorting layer, ant a2(s) search for a path of
Figure FDA0002783691790000042
And according to the search path length Lmax2Length update scheduling scheme if e2＜O2Otherwise, the process goes to step SS2, otherwise, the process goes to step SS 5.
Step SS 5: if the optimal extreme value of 5 sub-optimization is continuously
Figure FDA0002783691790000043
If there is no change, go to step SS6 to introduce a catastrophic factor, which is changedIf the path is a new path, otherwise, the batch sorting ant colony optimization is finished after the iteration times are finished, and the step SS7 is switched to;
step SS 6: introducing a catastrophe factor, changing the concentration of pheromone on a path of a machine distribution layer, and suddenly reducing or reducing the hormone to 0, thereby searching the optimal path in the global range again;
step SS 7: determining an optimal path and a scheme searched by a machine distribution layer algorithm;
step SS 8: if D > D, the algorithm stops; otherwise, the step SS9 is carried out;
step SS 9: and introducing a catastrophe factor, changing the concentration of pheromone on the order batch layer path, and suddenly reducing or reducing the hormone to 0, so as to search the optimal path in the global scope again.
5. The mixed flow batch scheduling optimization method for sensor shop production according to claim 4, wherein the step SS1 specifically comprises: generating an ant a according to the number of the sensors contained in each order1Dividing the order into a set of quanta of batches Wi(i ═ 1,2, …, N), ant a1Randomly selecting a batch W from a batch setiBeginning to traverse, set ant a1The traversed task counter is q1Storing the traversed task as 1
Figure FDA0002783691790000051
Ant a1And then according to the state transition rule, never traversing the task set
Figure FDA0002783691790000052
To select the next task, q1＝q1+ 1; then repeating the steps to complete all the tasks in a traversing way, generating a batch task sequence scheme, and recording the sequence scheme1＝e1+1。
6. The mixed flow batch scheduling optimization method for sensor shop production according to claim 4, wherein the steps areSS2 includes: selecting a machine for available processing for each process of each batch, specifically comprising: initializing available capacity K of each process machinem(M is 1,2, …, M), one ant a corresponding to the procedure is generated2(s) and selecting the rule as ant a according to the machine state2(s) initializing a batch node selection machine, setting ant a1The traversed batch counter is q21, ant a2(s) at the q-th2Step, selectable machine set Mx,s(ii) a From set M according to state transition rulesx,sIf the machine is selected, record q2＝q2+1, updating the capacity of the machines of the process s, continuously searching for the selection of a processing machine for each batch of each process, thus obtaining the final machine allocation scheme, note e2＝e2+1。
7. The mixed flow batch scheduling optimization method for sensor shop production according to claim 4, wherein the step SS3 specifically comprises:
step SS 31: initializing a process s, and acquiring a machine set M of the process ssTo generate an ant a3(m) determining a sequence batch on machine m by a batch transfer rule corresponding to one machine in the set of machines;
firstly, a certain processing batch on a certain machine is randomly selected for each ant as a node for starting traversal: let process s equal to 1, generate an ant a3(m) corresponding to one machine in the machine set to obtain a task set F of all batches on the corresponding machinen(N-1, 2, … N), all ants a3(m) random slave batch task set FnSelecting a batch; let the already-visited batch counter q3Storing the traversed batch set as 1
Figure FDA0002783691790000061
Then the ant a3(m) a set of unsearched batches
Figure FDA0002783691790000062
Then from
Figure FDA0002783691790000063
Select a certain batch of nodes and record q3＝q3+ 1; continuously repeating the process will traverse all the batch nodes, thereby obtaining the batch sequence, and recording e3＝e3+1；
Step SS 32: a decoding process, wherein a batch sequence dispatching result of the machine set is obtained;
(1) if the initial process is s equal to 1, the path length Q is initializedmax3Let Qmax30; otherwise, entering (2);
(2) assembling M from machinessExtracting machine M, calculating the machine completion time according to the formulas (2.4) - (2.7), and updating the machine set Ms；
(3) If it is
Figure FDA0002783691790000064
Then Q is updatedmax3If, if
Figure FDA0002783691790000065
Feeding in (4), otherwise, directly feeding in (4);
(4) if it is
Figure FDA0002783691790000066
Step SS33 is entered; otherwise, returning to the step (2);
step SS 33: calculating ant a3(m (path length L of search device sequence)max3(ii) a To a3(m) locally searching the searched equipment sequence to obtain a neighborhood sequence, comparing the path lengths of the neighborhood sequence and the original sequence, and taking the shortest path L'max3(ii) a Judging whether all the machines in all the steps are scheduled, if S is less than S, then S is equal to S +1, and then the operation goes to step SS 31; otherwise, judging whether each ant has traveled all the nodes of the batch, if e3＜O3If yes, the ant still has no journey, and the step is switched to SS 31; otherwise, the step SS34 is carried out;
step SS 34: if the optimal extreme value of 5 sub-optimization is continuously
Figure FDA0002783691790000071
If no change exists, the step SS35 is switched to introduce a catastrophe factor to change the catastrophe factor to a new path, otherwise, the iteration times are finished to finish the batch sorting ant colony algorithm of the equipment, and the step SS4 is switched to;
step SS 35: introducing a catastrophe factor, changing the concentration of pheromone on the path of the batch sequencing layer, and suddenly reducing or reducing the hormone to 0, thereby re-searching the optimal path in the global range;
Figure FDA0002783691790000072
then
Figure FDA0002783691790000073
And updating the current optimal scheme; the pheromone updating method comprises the following steps: attenuation of pheromone on long path obtained by search to tauij(t+n)＝(1-ρ)·τij(t)+ΔτijConversely, the shortest time-to-completion pathway increases pheromones to
Figure FDA0002783691790000074
Wherein a is3(min) the shortest ant length is selected,
Figure FDA0002783691790000075
Figure FDA0002783691790000076
proceed to step SS 31.
8. The mixed flow batch scheduling optimization method for sensor shop production according to claim 4, wherein the step SS6 specifically comprises: if it is
Figure FDA0002783691790000077
Then
Figure FDA0002783691790000078
And updating the machine allocation plan; the pheromone updating method comprises the following steps: attenuation of pheromone on long path obtained by search to taukm(t+n)＝(1-ρ)·τkm(t)+Δτkm(ii) a Conversely, the shortest completion time pathway increases its pheromone to
Figure FDA0002783691790000079
Wherein a is3(min) the shortest ant length is selected,
Figure FDA00027836917900000710
Figure FDA00027836917900000711
proceed to step SS 2.
9. The mixed flow batch scheduling optimization method for sensor shop production according to claim 4, wherein the step SS7 specifically comprises: ant a1The searched path is
Figure FDA0002783691790000081
And according to the search path length Lmax1A length update scheduling scheme; to a1The path of the sequence is locally searched to obtain a neighborhood sequence, the path lengths of the neighborhood sequence and the original sequence are compared, and the shortest path L 'is taken'max1. If e1＜O1Then, go to step SS 1; otherwise, the process goes to step SS 8.
10. The mixed flow batch scheduling optimization method for sensor shop production according to claim 4, wherein the step SS9 specifically comprises: if it is
Figure FDA0002783691790000082
Then
Figure FDA0002783691790000083
And update batchesA scheme; the pheromone updating method comprises the following steps: attenuation of pheromone on long path obtained by search to tauni(t+n)＝(1-ρ)·τni(t)+Δτkm(ii) a Conversely, the shortest completion time pathway increases its pheromone to
Figure FDA0002783691790000084
Figure FDA0002783691790000085
Wherein a is1(min) the shortest ant length is selected,
Figure FDA0002783691790000086
proceed to step SS 1.
CN202011290641.7A 2020-11-18 2020-11-18 Mixed flow batch scheduling optimization method for sensor workshop production Pending CN112462704A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011290641.7A CN112462704A (en) 2020-11-18 2020-11-18 Mixed flow batch scheduling optimization method for sensor workshop production

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011290641.7A CN112462704A (en) 2020-11-18 2020-11-18 Mixed flow batch scheduling optimization method for sensor workshop production

Publications (1)

Publication Number Publication Date
CN112462704A true CN112462704A (en) 2021-03-09

Family

ID=74837545

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011290641.7A Pending CN112462704A (en) 2020-11-18 2020-11-18 Mixed flow batch scheduling optimization method for sensor workshop production

Country Status (1)

Country Link
CN (1) CN112462704A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113570118A (en) * 2021-07-06 2021-10-29 浙江工业大学 Workshop scheduling and analyzing method based on scheduling rule
CN115130766A (en) * 2022-07-07 2022-09-30 西南交通大学 Multi-workshop automobile mixed flow and production scheduling collaborative optimization method considering multiple process routes
CN115719108A (en) * 2022-11-03 2023-02-28 吉林师范大学 Resource symmetric distributed workshop comprehensive scheduling method

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106971235A (en) * 2017-02-16 2017-07-21 上海大学 A kind of flexible job shop Optimization Scheduling in batches that there is intermediate storage constraint
CN106970604A (en) * 2017-05-15 2017-07-21 安徽大学 Multi-target workpiece scheduling algorithm based on ant colony algorithm
CN107831745A (en) * 2017-11-09 2018-03-23 西南交通大学 A kind of flexible job shop inserts single action state method for optimizing scheduling
CN109765862A (en) * 2018-12-26 2019-05-17 同济大学 The sustainable dispatch control method of hybrid flowshop based on self-adapted genetic algorithm
CN109918771A (en) * 2019-03-05 2019-06-21 北京工业大学 The energy-saving distribution model of hybrid flow forge under a kind of more time factors
CN110909858A (en) * 2019-11-27 2020-03-24 安徽大学 Batch scheduling method based on double-pheromone ant colony algorithm
CN110942251A (en) * 2019-11-27 2020-03-31 安徽大学 Batch scheduling method based on joint ant colony algorithm
CN111459628A (en) * 2020-03-12 2020-07-28 重庆邮电大学 Spark platform task scheduling method based on improved quantum ant colony algorithm

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106971235A (en) * 2017-02-16 2017-07-21 上海大学 A kind of flexible job shop Optimization Scheduling in batches that there is intermediate storage constraint
CN106970604A (en) * 2017-05-15 2017-07-21 安徽大学 Multi-target workpiece scheduling algorithm based on ant colony algorithm
CN107831745A (en) * 2017-11-09 2018-03-23 西南交通大学 A kind of flexible job shop inserts single action state method for optimizing scheduling
CN109765862A (en) * 2018-12-26 2019-05-17 同济大学 The sustainable dispatch control method of hybrid flowshop based on self-adapted genetic algorithm
CN109918771A (en) * 2019-03-05 2019-06-21 北京工业大学 The energy-saving distribution model of hybrid flow forge under a kind of more time factors
CN110909858A (en) * 2019-11-27 2020-03-24 安徽大学 Batch scheduling method based on double-pheromone ant colony algorithm
CN110942251A (en) * 2019-11-27 2020-03-31 安徽大学 Batch scheduling method based on joint ant colony algorithm
CN111459628A (en) * 2020-03-12 2020-07-28 重庆邮电大学 Spark platform task scheduling method based on improved quantum ant colony algorithm

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
付治政;肖菁;张军;: "基于信息素调整的蚁群算法求解JSP问题" *
付治政;肖菁;张军;: "基于信息素调整的蚁群算法求解JSP问题", 计算机工程与设计, no. 02 *
宋代立等: "蚁群算法求解混合流水车间分批调度问题", 《计算机集成制造系统》 *
宋代立等: "蚁群算法求解混合流水车间分批调度问题", 《计算机集成制造系统》, no. 07, 15 July 2013 (2013-07-15), pages 1640 - 1647 *
熊伟清;周扬;魏平;: "具有灾变的动态蚁群算法" *
熊伟清;周扬;魏平;: "具有灾变的动态蚁群算法", 电路与系统学报, no. 06 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113570118A (en) * 2021-07-06 2021-10-29 浙江工业大学 Workshop scheduling and analyzing method based on scheduling rule
CN113570118B (en) * 2021-07-06 2024-06-18 浙江工业大学 Workshop scheduling and analyzing method based on scheduling rules
CN115130766A (en) * 2022-07-07 2022-09-30 西南交通大学 Multi-workshop automobile mixed flow and production scheduling collaborative optimization method considering multiple process routes
CN115130766B (en) * 2022-07-07 2024-05-31 西南交通大学 Multi-workshop automobile mixed flow scheduling collaborative optimization method considering multi-process route
CN115719108A (en) * 2022-11-03 2023-02-28 吉林师范大学 Resource symmetric distributed workshop comprehensive scheduling method

Similar Documents

Publication Publication Date Title
CN112462704A (en) Mixed flow batch scheduling optimization method for sensor workshop production
Hyun et al. A genetic algorithm for multiple objective sequencing problems in mixed model assembly lines
Pan et al. Effective metaheuristics for scheduling a hybrid flowshop with sequence-dependent setup times
CN111966050A (en) AMMAS-GA nesting algorithm-based dual-resource mold job shop scheduling optimization method
CN107590603B (en) Based on the dispatching method and system for improving change neighborhood search and differential evolution algorithm
Coello Coello A comprehensive survey of evolutionary-based multiobjective optimization techniques
CN113792924A (en) Single-piece job shop scheduling method based on Deep reinforcement learning of Deep Q-network
Qin et al. Integrated production and distribution scheduling in distributed hybrid flow shops
TWI633504B (en) Tree search-based scheduling method and an apparatus using the same
Sirovetnukul et al. The impact of walking time on U-shaped assembly line worker allocation problems
CN113805545B (en) Flexible flow shop combined scheduling rule generation method considering batch processing
CN109345091A (en) Complete vehicle logistics dispatching method and device, storage medium, terminal based on ant group algorithm
CN108279647B (en) A kind of two stages assembly flow-shop scheduling
Basseur et al. Climbing combinatorial fitness landscapes
CN109271320A (en) A kind of upper multiple target priorities of test cases sort method
CN115130789A (en) Distributed manufacturing intelligent scheduling method based on improved wolf optimization algorithm
Lin Fuzzy job-shop scheduling based on ranking level (/spl lambda/, 1) interval-valued fuzzy numbers
Gu et al. A discrete particle swarm optimization algorithm with adaptive inertia weight for solving multiobjective flexible job-shop scheduling problem
Zaied et al. Permutation flow shop scheduling problem with makespan criterion: literature review
CN110928261B (en) Distributed estimation scheduling method and system for distributed heterogeneous flow shop
CN116466659A (en) Distributed assembly flow shop scheduling method based on deep reinforcement learning
CN115935616A (en) Multi-objective optimization method for scheduling of sequence-dependent flow shop groups of consistent batches
CN113515097B (en) Two-target single machine batch scheduling method based on deep reinforcement learning
Rifai et al. Reentrant FMS scheduling in loop layout with consideration of multi loading-unloading stations and shortcuts
Cao et al. An adaptive multi-strategy artificial bee colony algorithm for integrated process planning and scheduling

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20210309

RJ01 Rejection of invention patent application after publication