CN102938102A - Cross-operation unit scheduling method with batching machine - Google Patents

Abstract

The invention relates to a cross-operation unit scheduling method with a batching machine. The cross-operation unit scheduling method comprises the following steps of: 1, defining three information elements with different structures; 2, initializing the information elements; 3, assigning each process before a batching process of each part to a machine according to a machining order; 4, ordering the processes of each machine according to a time order; 5, batching and scheduling the parts; 6, assigning each process of each part after the batching process to the machine according to the machining order; 7, ordering the processes of each machine according to the time order; 8, updating the information elements according to a obtained solution; and 9, if a circulation frequency reaches an upper limit, or an optimal solution is unchanged for continuous times, finishing; and otherwise, turning to the step 2. With the adoption of the cross-operation unit scheduling method provided by the invention, the part cross-unit scheduling problem in a production process can be solved; and the integrated scheduling of a batching process and a non-batching process in the production process can be treated, and the operation efficiency can be ensured.

Description

A kind ofly stride the operation unit dispatching method with batch processor

Technical field

The present invention relates to a kind of dispatching method of manufacturing system, particularly a kind of with batch processor stride the operation unit dispatching method, belong to advanced production control and the optimizing scheduling field of making.

Background technology

Cell Manufacture System (Cellular Manufacturing System, CMS) is that the typical case of group technology (Group Technology, GT) in the field of manufacturing uses, and has embodied the philosophic theory of lean production.In CMS, machine is divided into groups according to the similarity of part process, form the relatively independent unit of working ability, the production run of one or more part families can be finished in each unit.Yet in actual production and since product day by day in variation and the unit productive capacity limited, the situation that exists some roads part operation of some part to process at the machine of other unit.Simultaneously, on the other hand, consider the reasons such as economics of production, budget and space constraint, for buying additional machine and infeasible in each unit.Because this class machine is usually expensive, startup once expends larger, considers the reasons such as economics of production, budget and space constraint, can only place it in certain specific unit.These equipment are as scarce resource, often can process a plurality of part families part (hereinafter this equipment is called share close reset standby, Critical Shared Machine, CSM).In this case, need to stride the peculiar part (Exceptional parts, EP) that just can finish a plurality of unit, the transfer between the unit has just formed strides unit branch problem (inter-cell move).Stride the unit transfer and cause each unit independently not dispatch, need the collaborative processing sequence that arranges part between the unit.

As far back as the early 1990s in last century, Garza and Smunt just point out to be difficult to avoid owing to striding the unit transfer, desirable CMS will be difficult to carry out, must quantitative test stride the unit and shift the impact that production system is produced, but most of research concentrates on always and how to carry out unit structure and how to carry out on the problem such as unit inner management.Until in recent years, along with the enforcement of CMS gradually deeply and meet difficulty, stride this problem of cell scheduling and just begin to be concerned, correlative study can be divided into strides the flowing water unit and strides two types of operation units.

Stride flowing water cell scheduling aspect, Yang and Liao consider at the most mobile transfer the between two unit of part, and only move and shift once.The shortest in target take flow-time, adopt branch and bound method and heuritic approach, solve the processing sequence of operation in each unit.Yet, along with plant configuration constantly changes to the unit manufacturing, strides the cell moving transfer and be popular tendency, and become increasingly complex, so the mobile branch problem of part between two above unit more merits attention.The people such as Solimanpur have considered the mobile branch problem of peculiar part between two above unit, take the makespan minimum as target, adopt heuritic approach to solve in two steps the cell scheduling problem of striding.The people such as Golmohammadi have considered mobile transfer the between the unit, adopt the dispatching sequence of part in the integrated solution part family of improved ElectroMagnetism-like (EM-like) algorithm dispatching sequence and each part family.The people such as Mosbah are to minimize makespan, and total completion date and standby time are that employing Extended Grate Deluge (EGD) algorithm and the heuritic approach of optimization aim solves the scheduling problem that has peculiar part in the Cell Manufacture System.In order to solve the Multi-Objective Scheduling of part in the unit and between the unit, the people such as Gholipour propose a kind of meta-heuristic algorithm based on scatter searching.

Stride the operation unit aspect, the people such as Tang use the scatter searching method to solve peculiar part and stride the cell scheduling problem.The people such as Elmi on the basis of people's problem models such as Tang, the permission part of can reentrying, namely the discontinuous operation of a part can be processed at same machine; Adopt the method for simulated annealing, the people's such as coded system and Tang scheme is identical, utilizes simultaneously proximity structure optimization finally to separate.The people such as Xiao have considered the situation that the part stochastic and dynamic arrives, and adopt Agent machinery of consultation based on pheromones to solve and stride the operation unit scheduling problem under the flexible path.

The problem model that the research to striding the cell scheduling problem in the research of above-mentioned list of references is considered is all according to based on traditional scheduling problem model, the scheduling model that production model (such as Flow Shop or scheduling model, job shop) is divided.Yet in actual production, Machine Type produces material impact to production scheduling equally, and this problem derives from the actual production process of vehicle key components and parts.

The process of part not only needs turnning and milling to dig the machining operations such as mill, also needs to anneal, mechanical property and the physical properties of heat treatment step to obtain to expect such as quenching, tempering, carburizing.Show have 35% the existing machine work order of part that heat treatment step is also arranged in the production run of vehicle key components and parts according to finding.In general, usually machine is added the separately consideration of scheduling of stage and heat treatment stages, reason is that the heat treatment time of parts adds the time much larger than machine usually.But our finding shows: see on the whole, though the machine work order time is shorter, but quantity is more, and therefore whole machine adds the time in stage and accounts for about 39% of production overall process, and heat treatment stages then accounts for about 42%, and the two is about the same; On the other hand, we find to reach thousands of minutes in investigation the process time that has a part of complicated machine work order at present, and these phenomenons are all so that machine adds stage and heat treatment stages most important optimization index---the time is mentioned in the same breath.Heat treating equipment only is placed in a certain discrete cell as a class CSM.Therefore different part families cause the access of heat treating equipment and stride the unit to shift the while also be transfer between uniprocessor (machining equipment) and the batch processor (heat treating equipment).The machine work order need to satisfy machine unique constraints (a machine synchronization can only be processed a part) and part unique constraints (a part synchronization can only by a machining), belongs to the solve job shop scheduling problems category; Heat treatment step need not satisfy the machine unique constraints, belongs to lot size scheduling problem category.Classical job-shop scheduling problem is a NP-hard problem, and the introducing of lot size scheduling problem makes it more complicated.The research of at present striding the cell scheduling problem from the angle analysis of different device types there is not yet achievement and delivers.

Summary of the invention

The objective of the invention is for the deficiencies in the prior art, striding under the unit collaboration mode, for the operation unit with batch processor finds efficient feasible dispatching method, to minimize completion date, maximization batch processor utilization factor and to minimize between non-batch processing operation and the batch processing operation stand-by period as target, guarantee the whole efficiently running of Cell Manufacture System.

The manufacturing system that the present invention considers is composed of multiple units, and some the machines that working ability is different are arranged in each unit, can finish the production of the part family of resemble process; There is and only has a batch processor in the system, be placed in one of them unit; Uniprocessor once can only be processed a part, and batch processor then can be processed a plurality of parts simultaneously; Manufacturing system also meets the following conditions:

1) all parts arrive constantly zero;

2) batch processor has certain space constraint, and the dimensions of part is known, is being no more than under the condition of space constraint, and the part of Same Part family can be processed in batch processor simultaneously;

3) process of part is comprised of the operation that multiple tracks has order constraint, and wherein each part has and only have a procedure to finish in batch processor, and the batch processing operation is not last procedure;

4) since between the unit machinery processing capacity overlap, cause part to have flexible path, but for the certain working procedure of certain part, but in each unit, have at the most a processing machine and process time known;

Be the maximal value of all part batch processing times in current batch the process time of 5) batch processing operation;

6) consider transfer time between the unit, and ignore transfer time in the unit, transfer time is different with the part classification because of distance between different units;

7) consider main setup time between the different part families, and ignore the less important setup time of Same Part family inside, namely when processing continuously the part of Same Part family on the uniform machinery, ignore setup time; Wherein main setup time is known and do not change with scheduling sequence and change;

8) buffer pool size before the batch processing equipment is enough large;

9) every machining workpiece all is the non-preemptive type of non-interruption;

10) do not consider the situation of mechanical disorder, shortage of raw materials and operating personnel's vacancy.

For the manufacturing system that satisfies above condition, the invention provides a kind of with batch processor stride the operation unit dispatching method, may further comprise the steps:

The 1st step: be defined as follows the index shown in the table and variable:

Table 1 index and variable

Simultaneously, the pheromones of three kinds of different structures of definition:

A) the pheromones structure during operation is assigned

Select in the process of machine in operation, the matrix that defines an O * M size represents pheromones, and wherein O represents the operation sum, and M represents machine sum, the element (O in the matrix _Ij, k) expression operation O _IjPheromone concentration corresponding to k processing on machine;

B) the pheromones structure in the Operation Sequencing

During Operation Sequencing on every machine, the matrix that defines M O * O size represents pheromones, and wherein O represents the operation sum, the element (O in m matrix _Ij, k) the upper operation O of expression machine m _IjK pheromone concentration that processing is corresponding on this machine;

C) the pheromones structure during batch processor operation group is criticized

In the process that batch processing operation group is criticized, the matrix of definition N * N size represents pheromones, and wherein N represents part sum, and element (i, j) expression part i and part j are at same batch of corresponding pheromone concentration;

Select a part to add existing batch at every turn from optional parts list when batch processing operation group is criticized, and batch number can't determine, therefore definition shown in the formula (1) is arranged:

Wherein, τ _i, b represents part i is added pheromone concentration corresponding to batch b, τ _{I, k}Expression part i and part k be at same batch of corresponding pheromone concentration, | B _b| expression B _bIn existing part number; Formula (1) is not if expression batch b be empty, and then pheromone concentration corresponding to part i adding batch b is, part i respectively with batch b in existing part same batch pheromone concentration sum, otherwise be definite value 1;

The 2nd step:

Carry out initialization, transfer distance, the part sum that must process and the technique information of each part between input machine information, dividing elements, unit, then according to following explanation initialization information element:

A) pheromones during the initialization operation is assigned

Wherein, τ _{I, j, m}Expression operation O _IjAt pheromone concentration corresponding to machine m processing, ε is the pheromone concentration initial value, is decided to be 0.01;

B) pheromones in the initialization Operation Sequencing

Wherein, τ on machine m _{M, i, j, k}Expression operation O _IjAt k the pheromone concentration that processing is corresponding, ε is the pheromone concentration initial value, is decided to be 0.01;

C) pheromones of initialization batch processor operation in batches

Wherein, τ _{I, k}Expression part i and part j are at the same batch of pheromone concentration that processing is corresponding, and ε is the pheromone concentration initial value, is decided to be 0.01;

The 3rd step:

According to processing sequence, the every procedure before the batch processing operation of each part is assigned to machine, namely be followed successively by the selected processing machine of every procedure for each part according to the processing sequence of operation, every selected probability of machine is:

${\mathrm{Pr}}_{i,j,m}=\frac{{\mathrm{\τ}}_{i,j,m}^{{\mathrm{\α}}_1}{\mathrm{\ρ}}_{i,j,m}^{{\mathrm{\β}}_1}}{{\mathrm{\Σ}}_{k=1}^M{\mathrm{\τ}}_{i,j,k}^{{\mathrm{\α}}_1}{\mathrm{\ρ}}_{i,j,k}^{{\mathrm{\β}}_1}}---\left(5\right)$

Wherein, Pr _{I, j, m}Expression operation O _IjThe probability that k processes on machine, ρ _{I, j, k}The heuristic information that expression is corresponding, α ₁, β ₁Represent respectively pheromone concentration, the shared weight of heuristic information;

Owing to having considered striding unit transfer time of part, therefore ρ _{I, j, k}Be defined as follows:

${\mathrm{\ρ}}_{i,j,k}=\frac{1}{P_{i,j,k}+{\mathrm{TT}}_i{D}_{m^{\′},k}}---\left(6\right)$

Wherein, P _{I, j, k}Expression O _IjProcess time on k on the machine, TT _iD _{M ', k}The machine m ' of expression part i unit distance transfer time and preceding working procedure processing amasss to the transfer distance of machine k is, i.e. transfer time corresponding to part i; Can be found out preferential Choice and process time and the less machine of sum transfer time when operation is assigned by the heuristic information formula;

Obtain operation behind the probability that every optional machine is processed, select at random certain this operation of machining with the roulette algorithm;

The 4th step:

According to time sequencing, with the per pass Operation Sequencing on each machine, namely on the basis that operation is assigned, determine processing sequencing and the zero-time of operation on every machine; Concrete grammar is:

For every machine, can dispatch the operation from it, according to probability shown in the following formula, select one procedure arrangement to process at the next one so that the roulette algorithm is random; Repeat this process, until all process steps all is scheduled;

${\mathrm{Pr}}_{m,i,j,k}=\frac{{\mathrm{\τ}}_{m,i,j,k}^{{\mathrm{\α}}_2}{\mathrm{\ρ}}_{m,i,j,k}^{{\mathrm{\β}}_2}}{{\mathrm{\Σ}}_{l=1}^O{\mathrm{\τ}}_{m,i,j,l}^{{\mathrm{\α}}_2}{\mathrm{\ρ}}_{m,i,j,l}^{{\mathrm{\β}}_2}}---\left(7\right)$

Wherein, Pr _{M, i, j, l}The upper O of expression machine m _IjAt the probability of l processing, ρ _{M, i, j, l}The heuristic information that expression is corresponding, α ₂, β ₂Represent respectively pheromone concentration, the shared weight of heuristic information;

The heuristic information of workpiece sequencing has considered the setup time of striding unit transfer time and machine of part, the setup time of striding unit transfer time and machine of part in fact can be overlapping simultaneously, namely in the transfer process of part, corresponding machine can begin to prepare for processing this part, therefore the heuristic information definition is suc as formula shown in (8):

${\mathrm{\ρ}}_{m,i,j,l}=\frac{1}{\max (f_{i,j-1}+{\mathrm{TT}}_i{D}_{m^{\′},m}{\mathrm{TE}}_{m,i,j}+l_{i,j}{\mathrm{ST}}_{i,m})-{\mathrm{TE}}_{m,i,j}}---\left(8\right)$

Wherein, f _{I, j-1}The time of the j-1 procedure process finishing of expression workpiece i, D _{M ', m}The transfer distance of expression from the machine m ' of a procedure on the processing work to machine m, l _{I, j}Affiliated part family: the l of part that is used for previous processing on the marking machine _{I, j}Workpiece and the workpiece i of=0 expression workpiece i previous processing on first processing or machine on the machine belong to Same Part family, l _{I, j}Workpiece and the workpiece i of previous processing do not belong to Same Part family, TE on=the 1 expression machine _{M, i, j}The upper O of expression machine m _IjFinish time of last procedure, max (f _{I, j-1}+ TT _iD _{I, j, m}, TE _{M, i, j}+ l _{I, j}ST _{I, j, m})-TE _{M, i, j}Represent that then a procedure finishes to O in the past _IjThe real time of beginning machining cost; By the heuristic information formula as can be known, when Operation Sequencing, priority scheduling can more early begin the operation of processing;

In the Operation Sequencing process, for any machine, it can dispatch operation is to be assigned in the operation of this machine, the set of the operation that the first operation of arbitrary part or last procedure have been sorted;

The 5th step:

The batch processing operation is criticized, dispatched to the part group, and the batch processing operation with which part of namely making a strategic decision is placed on same batch of batch processor and processes, and concrete grammar is:

Step1. select the time of arrival of part the earliest, add this batch;

Step2. upgrade candidate's part collection, candidate's part collection is defined as: if b is arbitrary batch, then its candidate's part collection is following set:

${\mathrm{CL}}_b=\left\{k\right|J_k=J_l,\&ForAll;l\&Element;B_b\mathrm{and}{S}_k<C_B-{\mathrm{\&Sigma;}}_{l\&Element;\mathrm{Bb}}{S}_l\mathrm{and\&Delta;}{\mathrm{WIS}}_b<0\}---\left(9\right)$

Be candidate's part collection of arbitrary batch, for all can not make with part Same Part family wherein and after adding this batch batch in the total size of part surpass the set of the part that batch processor capacity or batch waste and free space increase;

For a batch of b of batch processor, the waste of b and free space WIS _bBe this batch wasting space WS _bWith free space IS _bSum, for the scheduling solution S of a batch processing operation, WIS (S) represents waste and the free space of S, equals the WIS sum of all batches of S; WS and IS are defined as follows:

${\mathrm{WS}}_b=C_B\&CenterDot;P_b-{\mathrm{\&Sigma;}}_{i\&Element;B_b}{S}_i\&CenterDot;P_{\mathrm{iM}}---\left(10\right)$

IS _b＝C _B·(BS _b-BE _b-1) (11)

Wherein, BS _b, BE _bThe start and end time that represents respectively batch b, P _IMBe the batch processing operation process time of part i, and when b=0, make BE _B-1Be initial time,

Therefore have:

${\mathrm{WIS}}_b={\mathrm{WS}}_b+{\mathrm{IS}}_b=C_B\&CenterDot;P_b-\underset{i\&Element;B_b}{\mathrm{\&Sigma;}}{S}_i\&CenterDot;P_{\mathrm{iM}}+C_B\&CenterDot;({\mathrm{BS}}_b-{\mathrm{BE}}_{b-1})$

$=C_B\&CenterDot;({\mathrm{BE}}_b-{\mathrm{BE}}_{b-1})-{\mathrm{\&Sigma;}}_{i\&Element;B_b}{S}_i\&CenterDot;P_{\mathrm{iM}}---\left(12\right)$

Step3. concentrate the selected probability of part according to following formula calculated candidate part, then select one of them part to add this batch with the roulette algorithm;

Wherein, α ₃, β ₃Represent respectively pheromone concentration, the shared weight of heuristic information, Pr _{I, b}Expression part i adds the probability of batch b,

ΔWS _i，b＝WS _b′-WS _b＝C _B·(S _b′-S _b)+C _B·(P _b′-P _b)-S _iP _i (15)

If Step4. candidate's part collection is not empty, goes to Step2, otherwise continue to carry out Step5;

Criticize if Step5. also have part not yet to organize, go to Step1, continue to set up next batch, otherwise finish;

The 6th step:

According to processing sequence, the every procedure after the batch processing operation of each part is assigned to machine, method is with the 3rd step;

The 7th step:

According to time sequencing, with the per pass Operation Sequencing on each machine, method is with the 4th step;

The 8th step:

According to the solution that forms, the lastest imformation element, update rule is:

For each solution in the optimum solution set,

If a) operation O _IjBe assigned to machine m, then

τ _i，j，m＝(1-ρ)·τ _i，j，m+ρ·Δτ

B) if operation O _IjK processing on machine m, then

τ _{m，i，j，k}＝(1-ρ)·τ _{m，i，j，k}+ρ·Δτ

C) if part O _iAdd batch b, then

${\mathrm{\&tau;}}_{i,k}=(1-\mathrm{\&rho;})\&CenterDot;{\mathrm{\&tau;}}_{i,k}+\mathrm{\&rho;}\&CenterDot;\mathrm{\&Delta;\&tau;},\&ForAll;k\&Element;B_b$

Wherein, ρ represents the pheromones volatility, and Δ τ is the pheromones renewal amount:

$\mathrm{\&Delta;\&tau;}=Q\&CenterDot;({\mathrm{\&gamma;}}_1\frac{C_{\max }^l}{C_{\max }^g}+{\mathrm{\&gamma;}}_2\frac{R^l}{R^g}+{\mathrm{\&gamma;}}_3\frac{\mathrm{\&Delta;}{\mathrm{TW}}^l}{\mathrm{\&Delta;}{\mathrm{TW}}^g})$

Q is pheromones renewal amount factor of influence, γ ₁, γ ₂, γ ₃Be the weights of three optimization aim, represent respectively three optimization aim: the attention degree of stand-by period between maximum completion date, batch processor utilization factor, non-batch processing operation and the batch processing operation, and γ ₁+ γ ₂+ γ ₃=1,

Represent respectively total completion date that this scheduling is separated, batch processor utilization factor, the ratio of total waiting time and current optimum solution between non-batch processing operation and the batch processing operation;

The 9th step:

If cycle index reaches the upper limit, or the several times optimum solution is unchanged continuously, then finishes; Otherwise, turned for the 2nd step.

The present invention adopts the pheromones update method of max-min ant system (MMAS), in the process of upgrading, remains that pheromone concentration is in interval [τ _Min, τ _Max] interior (τ _MinAnd τ _MaxBe default definite value), the method is can avoiding method too early converges on locally optimal solution.

What participate in upgrading only is several locally optimal solutions during this circulates (being called optimal solution set down closes).The benefit of such update strategy is so that the scope of ant group hunting concentrate on more excellent solution near, and don't can be because only using the locally optimal solution in the circulation or the globally optimal solution lastest imformation is plain make speed of convergence excessively slow.

Beneficial effect

The present invention is directed to the solution that the operation unit scheduling problem proposes of striding of striding with batch processor, mainly contain following 3 beneficial effects:

(1) can solve in the production run part and stride the cell scheduling problem;

(2) the batch processing operation in can processing production process and the integrated scheduling of non-batch processing operation;

(3) operational efficiency is guaranteed.

Description of drawings

Fig. 1 is the solution process flow diagram.

Embodiment

The below specifies preferred implementation of the present invention.

The present embodiment according to the execution in step in the summary of the invention realize that the present invention proposes based on ant colony optimization algorithm with batch processor stride the operation unit dispatching algorithm, as shown in Figure 1.

The present embodiment is carried out following test simulation:

The Cell Manufacture System that the present invention is used for analog simulation arranges as follows: be provided with respectively 20 non-batch processors and 1 batch processor, dividing elements is as shown in table 2 under each machine.1 ~ 20 index is used in non-batch processor, and 21 index are used in batch processor.

Table 2 dividing elements

Note: optional machine index M1-20 represents non-batch processor in the table, and B21 represents batch processor.

The manufacturing process route of part is flexible, namely for same procedure, have the machine at least more than one unit can process this procedure, and the corresponding processing time of different machines is generally not identical.Problem model for the present invention proposes is provided with respectively 36 parts, and each part has different technological processes, and according to assumed condition, each part has and only has the batch processing operation one, and all the other operations are non-batch processing operation, and the batch processing operation is not the first operation.

But the handling machine that every procedure of emulation experiment data is corresponding generates by random algorithm.For non-batch processing operation, generate at random each 1 ~ 5 machine, lay respectively at different units.The size of part is chosen between 20 ~ 40 at random, and the process number of each part is chosen between 5 ~ 19 at random.The displacement of different parts between the unit is different, and displacement is chosen between 6 ~ 50 at random, owing to the part that belongs to different part families was chosen between 1 ~ 10 at random in the setup time that same machine switching produces.Choose at random the process time of the non-batch processing operation of per pass between 2 ~ 50, and choose at random the process time of batch processing operation between 100 ~ 200.

The environment of emulation experiment is as follows:

(1) operating system: Microsoft Windows732 position

(2)CPU：Intel Core 2Duo CPU T64002.00GHz

(3) internal memory: 2G

(4) development environment: Microsoft Visual Studio2010

(5) development language: C#

Parameter in the ACO algorithm has very large impact to its performance.Make parameter alpha ₁=α ₂=α ₃=1, by adjusting different β ₁, β ₂, β ₃Value reflection pheromone concentration and the shared weight of heuristic information.In addition, pheromones volatile ratio ρ, pheromones renewal amount factor of influence Q, pheromone concentration maximal value τ _MaxAlso the performance of ACO there is in various degree impact etc. parameter, therefore hereinafter described tests the reasonable value of determining parameter.Because the Q value is only relevant with the renewal amount of pheromone concentration, therefore only need determine Q and τ _MaxThe reasonable value of ratio gets final product, therefore when experiment with τ _MaxBe decided to be constant 5, to simplify experimentation.Need in the algorithm in this paper that span sees Table 3 in the parameter determined and the emulation experiment.

The experimental design of table 3 parameters simulation

Parameter	Scope
		β 1＝β 2	(0.2，0.5，1，2)
β 3	(0.01，0.1，0.5，1)
		Q/τ_max	(0.1，0.2，0.4)
ρ	(0.1，0.2，0.4)

For the parameter setting in the table 3, carried out contrast test during emulation.Because 3 different regulation goals of definition in the problem model, and wherein delivery date be one comparatively common and important in the production reality, therefore, when emulation experiment is carried out, the weight of delivery date, batch processor utilization factor, total waiting time is set to respectively 0.7,0.2,0.1, when evaluation algorithms uses the resulting solution of different parameters good and bad, use the weighted sum of the rank of value in all solutions of separating three corresponding regulation goals as standard.Experimental result (limit by length, only list the preceding front 20 groups of data of weighting rank herein) as shown in table 4:

Table 4 parameters simulation experimental result

Can draw β from experimental result ₁=2, β ₃=0.01, ρ=0.2, during Q/ τ _ max=0.2, the best performance of the solution that obtains.

Therefore, the user realizes when of the present invention, and suggestion is carried out following setting to parameter:

Parameter	Span	Recommended value
			α 1＝α 2＝α 3	1	1
β 1＝β 2	(0，4)	2
			β 3	(0，4)	0.01
ρ	(0,1)	0.2
			τ_max	(1，10)	5
Q	(0,τ_max)	0.2*τ_max

The Performance Ratio of the present embodiment in, Algorithm Performance is based on above-mentioned recommended value and obtains.

Emulation experiment adopt candidate's part collection rules different in the table 5 with (or) combination of different heuristic rule carries out contrast test and (hereinafter uses the CACO method, Combinational Ant Colony Optimization, the algorithm that acute pyogenic infection of finger tip is proposed by the invention).

When the batch processing operation is dispatched, adopt the concept of candidate's part collection, and used the heuristic information of WS as the ACO algorithm; When the scheduling of other operations, adopt the processing time and stride the formula that combines unit transfer time, setup time as heuristic information.Performance for checking CACO method, contrived experiment is considered 2 kinds of different candidate's part collection rules and group batch heuristic information, 1 kind of operation is assigned heuristic information, 1 kind of Operation Ordering Heuristics formula information makes up with the CACOI-CACOIV algorithm respectively, and above-mentioned 11 kinds of combined situation and algorithm CACO in this paper are compared.Specifically as shown in table 5.

The experimental design of table 5 contrast simulation

The below analyzes the result of emulation experiment with three performance index of total waiting time between maximum completion date, batch processor utilization factor, non-batch processing operation and the batch processing operation.In order to carry out intuitively the Comprehensive Correlation of three performance index, calculate respectively the Gap value that other contrast combinations are compared with the CACO method on three performance index during analysis, and according to 0.7,0.2,0.1 weight calculation weighting Gap value, and with the index of this Gap value as the evaluation algorithms performance.Computing formula is as follows:

${\mathrm{Gap}}_i^M=\frac{{\mathrm{Makespan}}_i-{\mathrm{Makespan}}_{\mathrm{CACO}}}{{\mathrm{Makespan}}_{\mathrm{CACO}}}$

${\mathrm{Gap}}_i^U=\frac{{\mathrm{UseRate}}_{\mathrm{CACO}}-{\mathrm{UseRate}}_i}{{\mathrm{UseRate}}_{\mathrm{CACO}}}$

${\mathrm{Gap}}_i^W=\frac{{\mathrm{WaitTime}}_i-{\mathrm{WaitTime}}_{i_{\mathrm{CACO}}}}{{\mathrm{WaitTime}}_{i_{\mathrm{CACO}}}}$

${\mathrm{Gap}}_i=0.7\&CenterDot;{\mathrm{Gap}}_i^M+0.2\&CenterDot;{\mathrm{Gap}}_i^R+0.1\&CenterDot;{\mathrm{Gap}}_i^W$

In the formula, between Makespan, UseRate, the maximum completion date of three performance index of WaitTime difference acute pyogenic infection of finger tip, batch processor utilization factor, non-batch processing operation and the batch processing operation,

Represent respectively the Gap value that a certain combination is compared with the CACO method on three performance index, Gap _iBe its weighted value.

Can be found out by above-mentioned the simulation experiment result, in the present embodiment, adopt FB still can obviously promote the batch processor utilization factor as candidate's part collection, but can have a strong impact on all the other two performance index, weighting Ga p is many 10%～30%, poor-performing.

The completion date that the combination of CACOI+WIS+CACOIII+CACOIV solves slightly is better than the solution of CACO, but the batch processor utilization factor has obvious decline, simultaneously, also has several combination, approach or slightly be better than CACO in a certain performance index, but on all the other indexs, all have certain gap.Can reach a conclusion in this case, candidate's part collection strategy that the CACO algorithm adopts and heuristic information overall performance are optimum.

It should be understood that present embodiment is instantiation of the invention process, should not be the restriction of protection domain of the present invention.In the situation that do not break away from spirit of the present invention and scope, modification or the change of foregoing being carried out equivalence all should be included within the present invention's scope required for protection.

Claims (2)

1. One kind with batch processor stride the operation unit dispatching method, may further comprise the steps:

   The 1st step: be defined as follows the index shown in the table and variable:

   Table 1 index and variable

   

   Simultaneously, the pheromones of three kinds of different structures of definition:

   A) the pheromones structure during operation is assigned

   Select in the process of machine in operation, the matrix that defines an O * M size represents pheromones, and wherein O represents the operation sum, and M represents machine sum, the element (O in the matrix _Ij, k) expression operation O _IjPheromone concentration corresponding to k processing on machine;

   B) the pheromones structure in the Operation Sequencing

   During Operation Sequencing on every machine, the matrix that defines M O * O size represents pheromones, and wherein O represents the operation sum, the element (O in m matrix _Ij, k) the upper operation O of expression machine m _IjK pheromone concentration that processing is corresponding on this machine;

   C) the pheromones structure during batch processor operation group is criticized

   In the process that batch processing operation group is criticized, the matrix of definition N * N size represents pheromones, and wherein N represents part sum, and element (i, j) expression part i and part j are at same batch of corresponding pheromone concentration;

   Select a part to add existing batch at every turn from optional parts list when batch processing operation group is criticized, and batch number can't determine, therefore definition shown in the formula (1) is arranged:

   

   Wherein, τ _{I, b}Expression adds pheromone concentration corresponding to batch b, τ with part i _{I, k}Expression part i and part k be at same batch of corresponding pheromone concentration, | B _b| expression B _bIn existing part number; Formula (1) is not if expression batch b be empty, and then pheromone concentration corresponding to part i adding batch b is, part i respectively with batch b in existing part same batch pheromone concentration sum, otherwise be definite value 1;

   The 2nd step:

   Carry out initialization, transfer distance, the part sum that must process and the technique information of each part between input machine information, dividing elements, unit, then according to following explanation initialization information element:

   A) pheromones during the initialization operation is assigned

   

   Wherein, τ _{I, j, m}Expression operation O _IjAt pheromone concentration corresponding to machine m processing, ε is the pheromone concentration initial value, is decided to be 0.01;

   B) pheromones in the initialization Operation Sequencing

   

   Wherein, τ on machine m _{M, i, j, k}Expression operation O _IjAt k the pheromone concentration that processing is corresponding, ε is the pheromone concentration initial value, is decided to be 0.01;

   C) pheromones of initialization batch processor operation in batches

   

   Wherein, τ _{I, k}Expression part i and part j are at the same batch of pheromone concentration that processing is corresponding, and ε is the pheromone concentration initial value, is decided to be 0.01;

   The 3rd step:

   According to processing sequence, the every procedure before the batch processing operation of each part is assigned to machine, namely be followed successively by the selected processing machine of every procedure for each part according to the processing sequence of operation, every selected probability of machine is:

   ${\mathrm{Pr}}_{i,j,m}=\frac{{\mathrm{\&tau;}}_{i,j,m}^{{\mathrm{\&alpha;}}_1}{\mathrm{\&rho;}}_{i,j,m}^{{\mathrm{\&beta;}}_1}}{{\mathrm{\&Sigma;}}_{k=1}^M{\mathrm{\&tau;}}_{i,j,k}^{{\mathrm{\&alpha;}}_1}{\mathrm{\&rho;}}_{i,j,k}^{{\mathrm{\&beta;}}_1}}---\left(5\right)$

   Wherein, Pr _{I, j, m}Expression operation O _IjThe probability that k processes on machine, ρ _{I, j, k}The heuristic information that expression is corresponding, α ₁, β ₁Represent respectively pheromone concentration, the shared weight of heuristic information;

   Owing to having considered striding unit transfer time of part, therefore ρ _{I, j, k}Be defined as follows:

   ${\mathrm{\&rho;}}_{i,j,k}=\frac{1}{P_{i,j,k}+{\mathrm{TT}}_i{D}_{m^{\&prime;},k}}---\left(6\right)$

   Wherein, P _{I, j, k}Expression O _IjProcess time on k on the machine, TT _iD _{M ', k}The machine m ' of expression part i unit distance transfer time and preceding working procedure processing amasss to the transfer distance of machine k is, i.e. transfer time corresponding to part i; Can be found out preferential Choice and process time and the less machine of sum transfer time when operation is assigned by the heuristic information formula;

   Obtain operation behind the probability that every optional machine is processed, select at random certain this operation of machining with the roulette algorithm;

   The 4th step:

   According to time sequencing, with the per pass Operation Sequencing on each machine, namely on the basis that operation is assigned, determine processing sequencing and the zero-time of operation on every machine; Concrete grammar is:

   For every machine, can dispatch the operation from it, according to probability shown in the following formula, select one procedure arrangement to process at the next one so that the roulette algorithm is random; Repeat this process, until all process steps all is scheduled;

   ${\mathrm{Pr}}_{m,i,j,k}=\frac{{\mathrm{\&tau;}}_{m,i,j,k}^{{\mathrm{\&alpha;}}_2}{\mathrm{\&rho;}}_{m,i,j,k}^{{\mathrm{\&beta;}}_2}}{{\mathrm{\&Sigma;}}_{l=1}^O{\mathrm{\&tau;}}_{m,i,j,l}^{{\mathrm{\&alpha;}}_2}{\mathrm{\&rho;}}_{m,i,j,l}^{{\mathrm{\&beta;}}_2}}---\left(7\right)$

   Wherein, Pr _{M, i, j, l}The upper O of expression machine m _IjAt the probability of l processing, ρ _{M, i, j, l}The heuristic information that expression is corresponding, α ₂, β ₂Represent respectively pheromone concentration, the shared weight of heuristic information;

   The heuristic information of workpiece sequencing has considered the setup time of striding unit transfer time and machine of part, the setup time of striding unit transfer time and machine of part in fact can be overlapping simultaneously, namely in the transfer process of part, corresponding machine can begin to prepare for processing this part, therefore the heuristic information definition is suc as formula shown in (8):

   ${\mathrm{\&rho;}}_{m,i,j,l}=\frac{1}{\max (f_{i,j-1}+{\mathrm{TT}}_i{D}_{m^{\&prime;},m}{\mathrm{TE}}_{m,i,j}+l_{i,j}{\mathrm{ST}}_{i,m})-{\mathrm{TE}}_{m,i,j}}---\left(8\right)$

   Wherein, f _{I, j-l}The time of the j-1 procedure process finishing of expression workpiece i, D _{M ', m}The transfer distance of expression from the machine m ' of a procedure on the processing work to machine m, l _{I, j}Affiliated part family: the l of part that is used for previous processing on the marking machine _{I, j}Workpiece and the workpiece i of=0 expression workpiece i previous processing on first processing or machine on the machine belong to Same Part family, l _{I, j}Workpiece and the workpiece i of previous processing do not belong to Same Part family, TE on=the 1 expression machine _{M, i, j}The upper O of expression machine m _IjFinish time of last procedure, max (f _{I, j-1}+ T T _iD _{I, j, m}, TE _{M, i, j}+ l _{I, j}ST _{I, j, m})-TE _{M, i, j}Represent that then a procedure finishes to O in the past _IjThe real time of beginning machining cost; By the heuristic information formula as can be known, when Operation Sequencing, priority scheduling can more early begin the operation of processing;

   In the Operation Sequencing process, for any machine, it can dispatch operation is to be assigned in the operation of this machine, the set of the operation that the first operation of arbitrary part or last procedure have been sorted;

   The 5th step:

   The batch processing operation is criticized, dispatched to the part group, and the batch processing operation with which part of namely making a strategic decision is placed on same batch of batch processor and processes, and concrete grammar is:

   Step1. select the time of arrival of part the earliest, add this batch;

   Step2. upgrade candidate's part collection, candidate's part collection is defined as: if b is arbitrary batch, then its candidate's part collection is following set:

   ${\mathrm{CL}}_b=\left\{k\right|J_k=J_l,\&ForAll;l\&Element;B_b\mathrm{and}{S}_k<-C_B-{\mathrm{\&Sigma;}}_{l\&Element;B_b}{S}_l\mathrm{and\&Delta;WI}{S}_b<0\}---\left(9\right)$

   Be candidate's part collection of arbitrary batch, for all can not make with part Same Part family wherein and after adding this batch batch in the total size of part surpass the set of the part that batch processor capacity or batch waste and free space increase;

   For a batch of b of batch processor, the waste of b and free space WIS _bBe this batch wasting space WS _bWith free space IS _bSum, for the scheduling solution S of a batch processing operation, WIS (S) represents waste and the free space of S, equals the WIS sum of all batches of S; WS and IS are defined as follows:

   ${\mathrm{WS}}_b=C_B\&CenterDot;P_b-{\mathrm{\&Sigma;}}_{i\&Element;B_b}{S}_i\&CenterDot;P_{\mathrm{iM}}---\left(10\right)$

   IS _b＝C _B·(BS _b-BE _b-1) (11)

   Wherein, BS _b, BE _bThe start and end time that represents respectively batch b, P _IMBe the batch processing operation process time of part i, and when b=0, make BE _B-1Be initial time,

   Therefore have:

   ${\mathrm{WIS}}_b={\mathrm{WS}}_b+{\mathrm{IS}}_b=C_B\&CenterDot;P_b-\underset{i\&Element;B_b}{\mathrm{\&Sigma;}}{S}_i\&CenterDot;P_{\mathrm{iM}}+C_B\&CenterDot;({\mathrm{BS}}_b-{\mathrm{BE}}_{b-1})$

   $=C_B\&CenterDot;({\mathrm{BE}}_b-{\mathrm{BE}}_{b-1})-{\mathrm{\&Sigma;}}_{i\&Element;B_b}{S}_i\&CenterDot;P_{\mathrm{iM}}---\left(12\right)$

   Step3. concentrate the selected probability of part according to following formula calculated candidate part, then select one of them part to add this batch with the roulette algorithm;

   

   Wherein, α ₃, β ₃Represent respectively pheromone concentration, the shared weight of heuristic information, Pr _{I, b}Expression part i adds the probability of batch b,

   

   ΔWS _i，b＝WS _b′-WS _b＝C _B·(S _b′-S _b)+C _B·(P _b′-P _b)-S _iP _i (15)

   If Step4. candidate's part collection is not empty, goes to Step2, otherwise continue to carry out Step5;

   Criticize if Step5. also have part not yet to organize, go to Step1, continue to set up next batch, otherwise finish;

   The 6th step:

   According to processing sequence, the every procedure after the batch processing operation of each part is assigned to machine, method is with the 3rd step;

   The 7th step:

   According to time sequencing, with the per pass Operation Sequencing on each machine, method is with the 4th step;

   The 8th step:

   According to the solution that forms, the lastest imformation element, update rule is:

   For each solution in the optimum solution set,

   If a) operation O _IjBe assigned to machine m, then

   τ _i，j，m＝(1-ρ)·τ _i，j，m+ρ·Δτ

   B) if operation O _IjK processing on machine m, then

   τ _{m，i，j，k}＝(1-ρ)·τ _{m，i，j，k}+ρ·Δτ

   C) if part O _iAdd batch b, then

   ${\mathrm{\&tau;}}_{i,k}=(1-\mathrm{\&rho;})\&CenterDot;{\mathrm{\&tau;}}_{i,k}+\mathrm{\&rho;}\&CenterDot;\mathrm{\&Delta;\&tau;},\&ForAll;k\&Element;B_b$

   Wherein, ρ represents the pheromones volatility, and Δ τ is the pheromones renewal amount:

   $\mathrm{\&Delta;\&tau;}=Q\&CenterDot;({\mathrm{\&gamma;}}_1\frac{C_{\max }^l}{C_{\max }^g}+{\mathrm{\&gamma;}}_2\frac{R^l}{R^g}+{\mathrm{\&gamma;}}_3\frac{{\mathrm{\&Delta;TW}}^l}{{\mathrm{\&Delta;TW}}_g})$

   Q is pheromones renewal amount factor of influence, γ ₁, γ ₂, γ ₃Be the weights of three optimization aim, represent respectively three optimization aim: the attention degree of stand-by period between maximum completion date, batch processor utilization factor, non-batch processing operation and the batch processing operation, and γ ₁+ γ ₂+ γ ₃=1,

   

   Represent respectively total completion date that this scheduling is separated, batch processor utilization factor, the ratio of total waiting time and current optimum solution between non-batch processing operation and the batch processing operation;

   The 9th step:

   If cycle index reaches the upper limit, or the several times optimum solution is unchanged continuously, then finishes; Otherwise, turned for the 2nd step;

   The scope that arranges of customer parameter is in above step:

   Parameter Span α ₁＝α ₂＝α ₃ 1 β ₁＝β ₂ (0，4) β ₃ (0，4) ρ (0,1) τ_max (1，10) Q (0,τ_max)
2. According to claim 1 a kind of with batch processor stride the operation unit dispatching method, it is characterized in that, according to customer parameter is set shown in the following table:

   Parameter Recommended value α ₁＝α ₂＝α ₃ 1

   β ₁＝β ₂ 2 β ₃ 0.01 ρ 0.2 τ_max 5 Q 0.2*τ_max

Images