CN109255484A - The discrete manufacturing recourses cooperative optimization method and system of data-driven - Google Patents

Abstract

The present invention provides the discrete manufacturing recourses cooperative optimization methods and system of a kind of data-driven.This method comprises: being analyzed by multifactor analysis of variance method historical production data, the core influence factor for influencing to manufacture efficiency is extracted, the net profit of all workpiece, i.e. fitness function are then calculated according to core influence factor.Later, the mechanism destroyed and reconstructed is executed to existing solution with reference to optimal solution, introduces variable neighborhood search algorithm, greediness is optimized with reference to iterative algorithm.The robustness and diversity of solution can be improved in the present invention, can effectively improve the production efficiency of the forging molding process of aluminium ingot, while can be generalized in complex product Discrete Manufacturing Process, reduces production cost.

Description

The discrete manufacturing recourses cooperative optimization method and system of data-driven

Technical field

The present invention relates to a kind of collaborations of the discrete manufacturing recourses of technical field of data processing more particularly to data-driven to optimizeMethod and system.

Background technique

With the fast development for manufacturing industry, traditional manufacturing cooperative scheduling needs to put into a large amount of manpower moneySource, degree of optimization rely heavily on the experience of personnel, and the trend that extensive, fining is manufactured is to conventionally manufactured modeBring new challenge.

Under the influence of many factors such as material characteristic, processing temperature and machine performance, Discrete Manufacturing Process is extremely complex,Such as in metal product forging and stamping, during actually generating of workpiece shaping, the discrete manufacturing recourses association with two stages degradation effectsIt is widely present with optimization problem, such as aluminium ingot needs to be pre-heated to certain temperature during machine-shaping, it is to be processed waitingDuring temperature decline, if processing when aluminium ingot temperature be higher than processing needed for temperature limit, can be processed；It is noThen, it needs to heat again and is restored to temperature limit, then processed again, it is therefore desirable to which it is excellent to be scheduled collaboration to aluminium ingot processingChange, reduce cost, improves production efficiency.

In the related technology, Barketau etc. (2008) considers the single machine for deteriorating workpiece with two stages and continuously criticizes scheduling and askTopic, it was demonstrated that it is strong np hard problem that target, which is the problem of minimizing completion date, under the situation.In addition, Leung etc. (2008) willSingle machine problem extends under multi-machine surroundings, it was demonstrated that even if the deterioration time limit of machine is all identical under multi-machine surroundings, the problem is stillIt is so strong np hard problem.

Summary of the invention

For the defects in the prior art, the present invention provides a kind of discrete manufacturing recourses of data-driven to cooperate with optimization sideMethod and system, for solving the problems, such as that resource distribution in the related technology is in the process strong np hard problem.

The embodiment of the invention provides a kind of discrete manufacturing recourses cooperative optimization methods of data-driven, comprising:

Step 1, using multifactor analysis of variance method analysis of history creation data, obtaining characterization influences the core manufacturedHeart factor；

Step 2, all workpiece and machine node are encoded using random secret key coding mode；

Step 3, its fitness value is calculated based on the initial solution generated at random, the initial solution is set as globally optimal solutionWith locally optimal solution；

Step 4, local search is carried out using based on the greedy local search algorithm with reference to iterative algorithm, redefines partOptimal solution selects the locally optimal solution redefined to update complete if the locally optimal solution redefined is better than globally optimal solutionOffice's optimal solution；

Step 5, x is enabled₁=x₂=1, i=1；

Step 6, it executes neighbour structure and changes N_i(x_i)；

Step 7, local search is carried out using based on the greedy local search algorithm with reference to iterative algorithm, obtains a partOptimal solution；If locally optimal solution is better than globally optimal solution, globally optimal solution is replaced using locally optimal solution；

Step 8, x is enabled_i=x_i+ 1, i=i+1；If i≤2, step 6 is returned to；Otherwise i=1 is enabled, goes to step 9；

Step 9, if x₂≤x_max, then return step 6；Otherwise, locally optimal solution and globally optimal solution are exported；

Step 10, it is decoded according to locally optimal solution and globally optimal solution, then aluminium ingot hot-working problem is given birth toProduce optimizing scheduling.

Optionally, carrying out coding to all workpiece and machine node using random secret key coding mode includes:

Random one length of generation is n+m-1, decimal array of the numerical value between 0-1；N and m respectively indicate piece countWith machine number；

The decimal array is mapped to the workpiece number array processed on each machine；

Each decimal in the decimal array is successively converted into integer from small to large, the integer is respectively 1,2 ... ...；

Using the integer greater than n as dividing mark, the subsequent workpiece of the dividing mark is distributed to other machines.

Optionally, local search is carried out with reference to iterative algorithm using based on greedy, obtaining a locally optimal solution includes:

1) maximum number of iterations ξ and reference based on the greedy local search mechanism with reference to iterative algorithm is arranged to covetGreedy iteration phase extracts the minimal amount q of position_minWith maximum value q_max；

2) initial solution is generated at random or a solution is generated by neighbour structure variation；The solution is that length is n+m-1, numberThe decimal array being worth between 0-1；

3) solution is decoded, the workpiece on every machine carries out batching according to LSPT rule, calculates fitness Fit；

4) value on any of solution position is randomly selected, the optimal location in solution is then inserted that value into；

5) q=q is enabled_min；

6) η=1 is enabled；

7) the value number that position is further extracted in setting is q, the value of q position in solution is randomly selected, after being extractedSolution becomes part solution π^partial, enable δ=1；

8) value of the δ extraction position [s] optimal into part solution is successively reinserted；PS_[s]It indicates on position [s]Value；

9) PS' is indicated and PS in the solution of part_[s]The same value, PS' is by optimal solution π^bestIt is divided into left and right two parts as ginsengIt examines；

10) δ=δ+1 is enabled, if δ > q, a disruption and recovery is completed, obtains a new solution, go to step10)；Otherwise return step 8)；

11) η=η+1 is enabled, if η≤ξ, return step 7)；Otherwise step 12) is gone to；

If 12) still do not obtain a preferably solution for iteration ξ times, q=q+1 is enabled, step 13) is gone to；Otherwise, it terminatesIteration；

If 13) q≤q_max, then return step 6)；Otherwise iteration is terminated.

Optionally, the step 9) includes:

9.1) a position is randomly choosed in the left-half of optimal solution, the value of the position is expressed as PS^L；

9.1.1) if part solution π^partialIn there are PS^LIt is worth, then PS in exchange part decomposition^LWithIf it does not exist, then it jumpsGo to step 9.1.3)；

9.1.2) if part solves π after exchange^partialFitness reduce, then jump to step 9.2)；Otherwise it carries out downOne step；

9.1.3 it) extractsValue, is then inserted into optimal position, so that part solution π^partialFitness mostIt is low；

9.2) PS is expressed as in the value that the right half part of optimal solution randomly chooses a position^R；

9.2.1) if part solution π^partialIn there are PS^RIt is worth, then PS in exchange part decomposition^RWithIf it does not exist, then it jumpsGo to step 9.2.3)；

9.2.2) if π is decomposed at exchange rear portion^partialFitness reduce, then jump to step 10)；Otherwise it carries out nextStep；

9.2.3 it) extractsValue is then inserted into optimal position, so that part solution π^partialFitness mostIt is low.

Optionally, carrying out local search based on the greedy local search algorithm with reference to iterative algorithm includes:

1) initial solution is generated at random or is changed by neighbour structure generates a solution π^initial；

2) it executes local search and finds a locally optimal solution, then replace initial solution π with it^initial；

If 3) π^initialIt is randomly generated, then enables π^incumbent=π^initial, π^best=π^initial, and q=q_min；Ifπ^initialIt is to be changed to generate by neighbour structure, then enables π^incumbent=π^initial, q=q_min；

4) θ=1 is enabled；

5) destructive process is executed, by π^incumbentQ position is randomly selected, remaining sequence generating unit is decomposedπ^partial；

6) restructuring procedure is executed, by reference to optimal solution π^best, π is successively decomposed into the value insertion section of q position^partial, obtainObtain a new solution π^new；

If 7) π^newFitness be lower than π^incumbentFitness, then by π^incumbentReplace with π^new；If π^newFitnessLower than π^bestFitness, then by π^bestReplace with π^new；

8) θ=θ+1 is enabled；If θ≤ξ, return step 5)；Otherwise step 9) is gone to；

If 9) do not obtain than preliminary examination solution π^initialPreferably solution, then enable q=q+1, go to step 10)；

If 10) q≤q_max, then return step 4)；Otherwise iteration is terminated.

Optionally, the step of calculating fitness value includes:

1) all workpiece on each machine are ranked up according to its core parameter non-decreasing；

2) machine M_jUpper piece count is expressed as n_j, the capacity of every batch of workpiece is c, and the workpiece on each machine is carried out groupBatch, comprising:

2.1) j=1 is set；

2.2) will beforeA workpiece composition a batch；

2.3) the preceding c workpiece of non-batching workpiece is formed a batch, if there are still non-batchings by the workpiece of non-batching if it existsWorkpiece, the step is continued to execute, until the workpiece whole batching on the machine；

2.4) machine M_jOn batch processed according to the sequencing of batching；

2.5) if j is less than the number of machine, j=j+1 is enabled, is gone to step 2.2)；Otherwise step 3) is carried out；

3) the completion date C of the last one batch on each machine is calculated_j, then net profit

4) the TNR value for calculating step 3) is as the fitness value of solution.

The embodiment of the invention also provides a kind of discrete manufacturing recourses of data-driven to cooperate with optimization system, comprising:

Central factor obtains module, for utilizing multifactor analysis of variance method analysis of history creation data, is characterizedInfluence the central factor manufactured；

Nodes encoding module, for being encoded using random secret key coding mode to all workpiece and machine node；

Fitness computing module is set the initial solution for being calculated its fitness value based on the initial solution generated at randomIt is set to globally optimal solution and locally optimal solution；

Optimal solution update module, for carrying out local search using based on the greedy local search algorithm with reference to iterative algorithmRope redefines locally optimal solution, if the locally optimal solution redefined is better than globally optimal solution, selects the office redefinedPortion's optimal solution updates globally optimal solution；

Initialization module, for enabling x₁=x₂=1, i=1；

Field changes module, for executing neighbour structure variation N_i(x_i)；

The optimal solution update module is also used to using based on the greedy local search algorithm carry out office for referring to iterative algorithmPortion's search, obtains a locally optimal solution；If locally optimal solution is better than globally optimal solution, replaced using locally optimal solution globalOptimal solution；

The initialization module is also used to enable x_i=x_i+ 1, i=i+1；If i≤2, the field changing pattern is triggeredBlock；Otherwise i=1 is enabled, optimal solution output module is triggered；

Optimal solution output module, in x₂≤x_maxWhen, then return to field variation module；Otherwise, locally optimal solution is exportedAnd globally optimal solution；

Then optimal solution decoder module adds aluminium ingot heat for being decoded according to locally optimal solution and globally optimal solutionWork problem carries out Optimization of Production Dispatching.

As shown from the above technical solution, it can be gone through by multifactor analysis of variance method to existing in the embodiment of the present inventionHistory creation data is analyzed, and can be extracted to influence and be manufactured the objective central factor of efficiency, so as to sufficiently benefitWith existing manufacturing data.Then the net profit of all workpiece, i.e. fitness function are calculated according to core influence factor, are madeThe foundation that model must be produced more rationally and is bonded practical with the calculating of fitness function.

Optimal solution being also referred in the present embodiment, the mechanism destroyed and reconstructed being executed to existing solution, introducing becomes neighborhood search and calculatesMethod optimizes greediness with reference to iterative algorithm, it is avoided to fall into locally optimal solution too early, improve the acquired robustness solved andDiversity effectively improves the production efficiency of the forging molding process of aluminium ingot, reduces production cost, while can be generalized to multipleIn miscellaneous product Discrete Manufacturing Process.

Optimization problem is cooperateed with by discrete resource during research aluminium ingot forging molding in the present embodiment, will be deteriorated two stagesEffect is introduced into the model of aluminium ingot forging molding, so that the model is more in line with the real process of aluminium ingot forging molding.

Detailed description of the invention

In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show belowThere is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only thisSome embodiments of invention for those of ordinary skill in the art without creative efforts, can be withOther attached drawings are obtained according to these figures.

Fig. 1 is that a kind of process of the discrete manufacturing recourses cooperative optimization method of data-driven provided in an embodiment of the present invention is shownIt is intended to；

Fig. 2 is that the discrete manufacturing recourses of data-driven provided in an embodiment of the present invention cooperate with the block diagram of optimization system.

Specific embodiment

Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, completeSite preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based onEmbodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every otherEmbodiment shall fall within the protection scope of the present invention.

Fig. 1 is the process signal of the discrete manufacturing recourses cooperative optimization method for the data-driven that one embodiment of the invention providesFigure.Referring to Fig. 1, the discrete manufacturing recourses cooperative optimization method of the data-driven includes step 1~step 10, in which:

Step 1, using multifactor analysis of variance method analysis of history creation data, the core for influencing to manufacture in table is obtainedHeart factor.

In this step, by the length of such as workpiece of the Multiple factors in historical production data, thickness of workpiece, workpiece temperature,The factors such as the normal processing speed of machine find the process time of the dependent variable such as workpiece in creation data as independent variable.SoVariance analysis is done to independent variable and dependent variable afterwards, find significantly affect the factor of dependent variable as the core for influencing to manufacture becauseElement.In this way, the present embodiment can filter out more smaller or without influence the influence of influence, calculation amount is reduced.

Step 2, all workpiece and machine node are encoded using random secret key coding mode.

In this step, 1) generate a length at random as n+m-1, decimal array of the numerical value between 0-1, n and m distinguish tableShow piece count and machine number.

2) the decimal array is mapped to the workpiece number array processed on each machine；

2.1) each decimal in the decimal array being successively converted into integer from small to large, the integer is respectively 1,2 ... ...；

2.1) using the integer greater than n as dividing mark, the subsequent workpiece of the dividing mark is distributed to other machines.

For example, there are 6 workpiece and 3 machines, decimal array (0.41,0.03,0.76,0.25,0.65,0.89,0.35,0.28) being mapped to integer array is (5,1,7,2,9,6,8,4,3), is expressed as workpiece 5 and workpiece 1 in First machineUpper processing, work 2 and workpiece 6 are processed on second machine, and workpiece 4 and workpiece 3 are processed on third platform machine.

Step 3, its fitness value is calculated based on the initial solution generated at random, the initial solution is set as globally optimal solutionWith locally optimal solution π^best。

In this step, calculate fitness value the step of include:

1) all workpiece on each machine are ranked up according to its core parameter non-decreasing；

2) machine M_jUpper piece count is expressed as n_j, the capacity of every batch of workpiece is c, and the workpiece on each machine is carried out groupBatch, comprising:

2.1) j=1 is set；

2.2) will beforeA workpiece composition a batch；

2.3) the preceding c workpiece of non-batching workpiece is formed a batch, if there are still non-batchings by the workpiece of non-batching if it existsWorkpiece, the step is continued to execute, until the workpiece whole batching on the machine；

2.4) machine M_jOn batch processed according to the sequencing of batching；

2.5) if j is less than the number of machine, j=j+1 is enabled, is gone to step 2.2)；Otherwise step 3) is carried out；

3) the completion date C of the last one batch on each machine is calculated_j, then net profit

4) the TNR value for calculating step 3) is as the fitness value of solution.

Step 4, local search is carried out using based on the greedy local search algorithm with reference to iterative algorithm, redefines partOptimal solution selects the locally optimal solution redefined to update complete if the locally optimal solution redefined is better than globally optimal solutionOffice's optimal solution.

In this step, carrying out local search based on the greedy local search algorithm with reference to iterative algorithm includes:

1) maximum number of iterations ξ and reference based on the greedy local search mechanism with reference to iterative algorithm is arranged to covetGreedy iteration phase extracts the minimal amount q of position_minWith maximum value q_max；

2) initial solution is generated at random or a solution is generated by neighbour structure variation；

3) solution is decoded, the workpiece on every machine carries out batching according to LSPT rule, calculates fitness Fit；

4) value on any of solution position is randomly selected, the optimal location in solution is then inserted that value into；

5) q=q is enabled_min；

6) η=1 is enabled；

7) the value number that position is further extracted in setting is q, the value of q position in solution is randomly selected, after being extractedSolution becomes part solution π^partial, enable δ=1；

8) value of the δ extraction position [s] optimal into part solution is successively reinserted；PS_[s]It indicates on position [s]Value；

9) PS' is indicated and PS in the solution of part_[s]The same value, PS' is by optimal solution π^bestIt is divided into left and right two parts as ginsengIt examines；

10) δ=δ+1 is enabled, if δ > q, a disruption and recovery is completed, obtains a new solution, go to step10)；Otherwise return step 8)；

11) η=η+1 is enabled, if η≤ξ, return step 7)；Otherwise step 12) is gone to；

If 12) still do not obtain a preferably solution for iteration ξ times, q=q+1 is enabled, step 13) is gone to；Otherwise, it terminatesIteration；

If 13) q≤q_max, then return step 6)；Otherwise iteration is terminated.

Then, it regard the better solution of output as locally optimal solution.If the locally optimal solution redefined is most better than the overall situationExcellent solution then selects the locally optimal solution redefined to update globally optimal solution.

Step 5, x is enabled₁=x₂=1, i=1；

Step 6, it executes neighbour structure and changes N_i(x_i)；

Step 7, local search is carried out using based on the greedy local search algorithm with reference to iterative algorithm, obtains a partOptimal solution；If locally optimal solution is better than globally optimal solution, globally optimal solution is replaced using locally optimal solution.

In this step, the step of obtaining locally optimal solution, is referring to step 4, and the step of updating globally optimal solution can also joinSee step 4, therefore not to repeat here.

Step 8, x is enabled_i=x_i+ 1, i=i+1；If i≤2, step 6 is returned to；Otherwise i=1 is enabled, goes to step 9.

Step 9, if x₂≤x_max, then return step 6；Otherwise, locally optimal solution and globally optimal solution are exported；

Step 10, it is decoded according to locally optimal solution and globally optimal solution, then aluminium ingot hot-working problem is given birth toProduce optimizing scheduling.

As shown from the above technical solution, it can be gone through by multifactor analysis of variance method to existing in the embodiment of the present inventionHistory creation data is analyzed, and can be extracted to influence and be manufactured the objective central factor of efficiency, so as to sufficiently benefitWith existing manufacturing data.Then the net profit of all workpiece, i.e. fitness function are calculated according to core influence factor, are madeThe foundation that model must be produced more rationally and is bonded practical with the calculating of fitness function.

Optimal solution being also referred in the present embodiment, the mechanism destroyed and reconstructed being executed to existing solution, introducing becomes neighborhood search and calculatesMethod optimizes greediness with reference to iterative algorithm, it is avoided to fall into locally optimal solution too early, improve the acquired robustness solved andDiversity effectively improves the production efficiency of the forging molding process of aluminium ingot, reduces production cost.

As it can be seen that optimization problem is cooperateed with by discrete resource during research aluminium ingot forging molding in the present embodiment, by two ranksSection degradation effects are introduced into the model of aluminium ingot forging molding, so that the model is more in line with the practical mistake of aluminium ingot forging moldingJourney.

Fig. 2 is that the discrete manufacturing recourses for the data-driven that one embodiment of the invention provides cooperate with the block diagram of optimization system.GinsengSee Fig. 2, the discrete manufacturing recourses collaboration optimization system of data-driven includes:

Central factor obtains module 201, for utilizing multifactor analysis of variance method analysis of history creation data, obtains tableSign influences the central factor manufactured；

Nodes encoding module 202, for being compiled using random secret key coding mode to all workpiece and machine nodeCode；

Fitness computing module 203, for calculating its fitness value based on the initial solution generated at random, by the initial solutionIt is set as globally optimal solution and locally optimal solution；

Optimal solution update module 204, for carrying out part using based on the greedy local search algorithm with reference to iterative algorithmSearch, redefines locally optimal solution, if the locally optimal solution redefined is better than globally optimal solution, what selection redefinedLocally optimal solution updates globally optimal solution；

Initialization module 205, for enabling x₁=x₂=1, i=1；

Field changes module 206, for executing neighbour structure variation N_i(x_i)；

The optimal solution update module 204, be also used to using based on the greedy local search algorithm with reference to iterative algorithm intoRow local search obtains a locally optimal solution；If locally optimal solution is better than globally optimal solution, replaced using locally optimal solutionGlobally optimal solution；

The initialization module 205, is also used to enable x_i=x_i+ 1, i=i+1；If i≤2, the field variation is triggeredModule；Otherwise i=1 is enabled, optimal solution output module is triggered；

Optimal solution output module 207, in x₂≤x_maxWhen, then return to field variation module；Otherwise, output is local mostExcellent solution and globally optimal solution；

Optimal solution decoder module 208, for being decoded according to locally optimal solution and globally optimal solution, then to aluminium ingot heatProcessing problems carry out Optimization of Production Dispatching.

It should be noted that data-driven provided in an embodiment of the present invention discrete manufacturing recourses collaboration optimization system with it is upperThe method of stating is one-to-one relationship, and the implementation detail of the above method is equally applicable to above system, and the embodiment of the present invention is no longerAbove system is described in detail.

In specification of the invention, numerous specific details are set forth.It is to be appreciated, however, that the embodiment of the present invention can be withIt practices without these specific details.In some instances, well known method, structure and skill is not been shown in detailArt, so as not to obscure the understanding of this specification.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention., rather than its limitations；To the greatest extentPipe present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: its according toSo be possible to modify the technical solutions described in the foregoing embodiments, or to some or all of the technical features intoRow equivalent replacement；And these are modified or replaceed, various embodiments of the present invention technology that it does not separate the essence of the corresponding technical solutionThe range of scheme should all cover within the scope of the claims and the description of the invention.

Claims (7)

1. a kind of discrete manufacturing recourses cooperative optimization method of data-driven characterized by comprising

Step 1, using multifactor analysis of variance method analysis of history creation data, obtain characterization influence the core manufactured becauseElement；

Step 2, all workpiece and machine node are encoded using random secret key coding mode；

Step 3, its fitness value is calculated based on the initial solution generated at random, the initial solution is set as globally optimal solution and officePortion's optimal solution；

Step 4, local search is carried out using based on the greedy local search algorithm with reference to iterative algorithm, redefines local optimumSolution selects the locally optimal solution redefined to update the overall situation most if the locally optimal solution redefined is better than globally optimal solutionExcellent solution；

Step 5, x is enabled₁=x₂=1, i=1；

Step 6, it executes neighbour structure and changes N_i(x_i)；

Step 7, local search is carried out using based on the greedy local search algorithm with reference to iterative algorithm, obtains a local optimumSolution；If locally optimal solution is better than globally optimal solution, globally optimal solution is replaced using locally optimal solution；

Step 8, x is enabled_i=x_i+ 1, i=i+1；If i≤2, step 6 is returned to；Otherwise i=1 is enabled, goes to step 9；

Step 9, if x₂≤x_max, then return step 6；Otherwise, locally optimal solution and globally optimal solution are exported；

Step 10, it is decoded according to locally optimal solution and globally optimal solution, production tune then is carried out to aluminium ingot hot-working problemDegree optimization.

2. discrete manufacturing recourses cooperative optimization method according to claim 1, which is characterized in that encoded using random secret keyMode carries out coding to all workpiece and machine node

Random one length of generation is n+m-1, decimal array of the numerical value between 0-1；N and m respectively indicate piece count and machineDevice number；

The decimal array is mapped to the workpiece number array processed on each machine；

Each decimal in the decimal array is successively converted into integer from small to large, the integer is respectively 1,2 ... ...；

Using the integer greater than n as dividing mark, the subsequent workpiece of the dividing mark is distributed to other machines.

3. discrete manufacturing recourses cooperative optimization method according to claim 1, which is characterized in that using based on greedy referenceIterative algorithm carries out local search, and obtaining a locally optimal solution includes:

1) the maximum number of iterations ξ based on the greedy local search mechanism with reference to iterative algorithm is set and is changed with reference to greedinessThe minimal amount q of position is extracted for the stage_minWith maximum value q_max；

2) initial solution is generated at random or a solution is generated by neighbour structure variation；The solution is that length is n+m-1, and numerical value existsDecimal array between 0-1；

3) solution is decoded, the workpiece on every machine carries out batching according to LSPT rule, calculates fitness Fit；

4) value on any of solution position is randomly selected, the optimal location in solution is then inserted that value into；

5) q=q is enabled_min；

6) η=1 is enabled；

7) setting further extract position value number be q, randomly select the value of q position in solution, the solution after being extracted atPartially to solve π^partial, enable δ=1；

8) value of the δ extraction position [s] optimal into part solution is successively reinserted；PS_[s]Indicate the value on position [s]；

9) PS' is indicated and PS in the solution of part_[s]The same value, PS' is by optimal solution π^bestIt is divided into left and right two parts as reference；

10) δ=δ+1 is enabled, if δ > q, a disruption and recovery is completed, obtains a new solution, go to step 10)；It is noThen return step 8)；

11) η=η+1 is enabled, if η≤ξ, return step 7)；Otherwise step 12) is gone to；

If 12) still do not obtain a preferably solution for iteration ξ times, q=q+1 is enabled, step 13) is gone to；Otherwise, termination changesGeneration；

If 13) q≤q_max, then return step 6)；Otherwise iteration is terminated.

4. discrete manufacturing recourses cooperative optimization method according to claim 3, which is characterized in that the step 9) includes:

9.1) a position is randomly choosed in the left-half of optimal solution, the value of the position is expressed as PS^L；

9.1.1) if part solution π^partialIn there are PS^LIt is worth, then PS in exchange part decomposition^LWithIf it does not exist, then it jumps toStep 9.1.3)；

9.1.2) if part solves π after exchange^partialFitness reduce, then jump to step 9.2)；Otherwise it carries out in next step；

9.1.3 it) extractsValue, is then inserted into optimal position, so that part solution π^partialFitness it is minimum；

9.2) PS is expressed as in the value that the right half part of optimal solution randomly chooses a position^R；

9.2.1) if part solution π^partialIn there are PS^RIt is worth, then PS in exchange part decomposition^RWithIf it does not exist, then it jumps toStep 9.2.3)；

9.2.2) if π is decomposed at exchange rear portion^partialFitness reduce, then jump to step 10)；Otherwise it carries out in next step；

9.2.3 it) extractsValue is then inserted into optimal position, so that part solution π^partialFitness it is minimum.

5. discrete manufacturing recourses cooperative optimization method according to claim 1, which is characterized in that refer to iteration based on greedyThe variation solved in the local search algorithm of algorithm includes:

1) initial solution is generated at random or is changed by neighbour structure generates a solution π^initial；

2) it executes local search and finds a locally optimal solution, then replace initial solution π with it^initial；

If 3) π^initialIt is randomly generated, then enables π^incumbent=π^initial, π^best=π^initial, and q=q_min；If π^initialIt is to be changed to generate by neighbour structure, then enables π^incumbent=π^initial, q=q_min；

4) θ=1 is enabled；

5) destructive process is executed, by π^incumbentQ position is randomly selected, remaining sequence generating unit decomposes π^partial；

6) restructuring procedure is executed, by reference to optimal solution π^best, π is successively decomposed into the value insertion section of q position^partial, obtain oneA new solution π^new；

If 7) π^newFitness be lower than π^incumbentFitness, then by π^incumbentReplace with π^new；If π^newFitness be lower thanπ^bestFitness, then by π^bestReplace with π^new；

8) θ=θ+1 is enabled；If θ≤ξ, return step 5)；Otherwise step 9) is gone to；

If 9) do not obtain than preliminary examination solution π^initialPreferably solution, then enable q=q+1, go to step 10)；

If 10) q≤q_max, then return step 4)；Otherwise iteration is terminated.

6. discrete manufacturing recourses cooperative optimization method according to claim 1, which is characterized in that calculate the step of fitness valueSuddenly include:

1) all workpiece on each machine are ranked up according to its core parameter non-decreasing；

2) machine M_jUpper piece count is expressed as n_j, the capacity of every batch of workpiece is c, and the workpiece on each machine is carried out batching, packetIt includes:

2.1) j=1 is set；

2.2) will beforeA workpiece composition a batch；

2.3) the preceding c workpiece of non-batching workpiece is formed a batch, if there are still the works of non-batching by the workpiece of non-batching if it existsPart continues to execute the step, until the workpiece whole batching on the machine；

2.4) machine M_jOn batch processed according to the sequencing of batching；

2.5) if j is less than the number of machine, j=j+1 is enabled, is gone to step 2.2)；Otherwise step 3) is carried out；

3) the completion date C of the last one batch on each machine is calculated_j, then net profit

4) the TNR value for calculating step 3) is as the fitness value of solution.

7. a kind of discrete manufacturing recourses of data-driven cooperate with optimization system characterized by comprising

Central factor obtains module, and for utilizing multifactor analysis of variance method analysis of history creation data, obtaining characterization influencesThe central factor of the manufacturing；

Nodes encoding module, for being encoded using random secret key coding mode to all workpiece and machine node；

The initial solution is set as by fitness computing module for being calculated its fitness value based on the initial solution generated at randomGlobally optimal solution and locally optimal solution；

Optimal solution update module, for carrying out local search, weight using based on the greedy local search algorithm with reference to iterative algorithmIt is new to determine locally optimal solution, if the locally optimal solution redefined is better than globally optimal solution, select the part redefined mostExcellent solution updates globally optimal solution；

Initialization module, for enabling x₁=x₂=1, i=1；

Field changes module, for executing neighbour structure variation N_i(x_i)；

The optimal solution update module is also used to carry out local search using based on the greedy local search algorithm with reference to iterative algorithmRope obtains a locally optimal solution；If locally optimal solution is better than globally optimal solution, global optimum is replaced using locally optimal solutionSolution；

The initialization module is also used to enable x_i=x_i+ 1, i=i+1；If i≤2, the field variation module is triggered；It is noI=1 is then enabled, optimal solution output module is triggered；

Optimal solution output module, in x₂≤x_maxWhen, then return to field variation module；Otherwise, output locally optimal solution and completeOffice's optimal solution；

Then optimal solution decoder module asks aluminium ingot hot-working for being decoded according to locally optimal solution and globally optimal solutionTopic carries out Optimization of Production Dispatching.