CN104751297A - Productivity allocation method for mixed-model production line - Google Patents

Abstract

The invention discloses a productivity allocation method for a mixed-model production line. The productivity allocation method includes the implementing steps of passing verification module permission information by a user and inputting product information of a motor production line into a control system; preprocessing the information to convert production data into specific productivity allocation models, user setting production line model parameters, control policy parameters and control system interaction protocols; performing operation on a production line productivity allocation and control module, wherein the production line productivity allocation and control module specifically comprises a production line buffer area size adjustment and control submodule, an empire-competition-algorithm-based optimized product batch division submodule and a product batch production sequence regulation and control submodule; making judgment on termination of buffer area optimization; allocating productivity of the production line and displaying control results; enabling the user to pass module permission and send the optimized results to a server so as to enable workshop management personnel to execute tasks according to the results. The productivity allocation method for the mixed-model production line has the advantages that the optimal sequence of workpiece batches is achieved, productivity allocation efficiency is improved, and productivity of the production line is further optimized.

Description

A kind of mix flow Productivity Allocation method

Technical field

The present invention relates to a kind of distribution method, particularly relate to a kind of mix flow Productivity Allocation method.

Background technology

In the face of present customization customization of individual character market, multi items mix flow is a kind of mode of production that current enterprise generally adopts, by changing the Productivity Allocation of mix flow, within a certain period of time, same production line can be produced the product of multiple different model, varying number.And in the production system of order-driven market, the key issue that the enterprise that how limited production capacity and Production requirement matched faces.Effective mix flow Productivity Allocation method and control system can not only improve production capacity and the production efficiency of enterprise, variation and the customer demand of enterprise fast responding market can also be made, show one's talent in the market competition of fierceness, thus win the larger market share, create more economic benefit.

The Productivity Allocation of mix flow and control problem relate to that production buffer district size controls, batch of orders divides and the subproblem such as Order Processing route optimization, are the combinatorial optimization problems that a class is very complicated.The mix flow Productivity Allocation method of current comparative maturity mainly comprises stochastic programming method, based on the capacity planning method and linear programming method etc. of input output model, all there is shot array problem in these methods, be difficult to be applicable to large-scale production system, and to " obstruction " occurred frequent in production line, (inter process buffer zone is full, preceding working procedure cannot continue processing) and " hunger " (inter process buffer zone is empty, later process does not have work pieces process) etc. dynamic event could not consider, and these events above all can produce material impact to the production capacity of production line, therefore existing method be not suitable for the production application of enterprise, also be difficult to obtain good production performance.

Kingdom's Competitive Algorithms be subject to imperialst state's warfare inspiration and by Atashpaz and Lucas in 2007 propose a kind of evolutional algorithm.This algorithm searches for from an initial population.Individuality in initial population is called as country, and they are divided into two classes: colony and suzerainty state, and each suzerainty state forms a kingdom with several colonies being attached to it.There is the colonial competition of contention mutually between each kingdom, the kingdom that strength is strong will obtain increasing colony, the colony that then will lose oneself gradually that strength is weak, until be destroyed.The strength of a kingdom depends on suzerain strength and colonial strength simultaneously.Net result after algorithm performs is that All Countries forms a kingdom, and algorithm stops search.There are some researches show, kingdom's Competitive Algorithms has good ability of searching optimum and speed of convergence efficiently for solving of combinatorial optimization problem, is the intelligent algorithm that the very applicable extensive combinatorial optimization problem of a kind of performance solves.

Summary of the invention

In order to solve the weak point existing for above-mentioned technology, the invention provides a kind of mix flow Productivity Allocation method.

In order to solve above technical matters, the technical solution used in the present invention is: a kind of mix flow Productivity Allocation method, and performing step is as follows:

Motor product of production line information, by authentication module License Info, is input to control system by browser by A, user;

B, information pre-processing.Transform that production data is concrete Productivity Allocation model, user arranges production line model parameter, control strategy parameter and control system interaction protocol.

C, production line Productivity Allocation and control module; Specifically comprise production line buffer size regulator module, kingdom's Competitive Algorithms optimizing product batch divides submodule and product batches production sequence regulator module.

C.1, production buffer district size control submodule is arranged

Equipment room buffer size initial value buffer is set ₀with the incremental change Δ of each circulation,

buffer _i＝buffer ₀+i*Δ

C.2, kingdom's Competitive Algorithms search;

The buffer size of equipment room calculated according to step 1 and the processing batch size of often kind of product after optimizing, set up integer programming model, calculate the optimum process-cycle, as the basis of each kingdom Competitive Algorithms ideal adaptation angle value, process-cycle after the Competitive Algorithms convergence of final kingdom, batch size corresponding with it and product batches processing sequence were as production capacity prioritization scheme as the rreturn value of step C.

C.2.1, kingdom's Competitive Algorithms initialization

C.2.1.1, algorithm parameter arrange, setting parameter, the country number N of set algorithm _pop, emperialist number N _imp, colony weight factor ξ, revolution rate r and maximum cycle N;

C.2.1.2, initial population is produced

Counter i=0 is set, at [1, product_num _i] interior repetition N _popsecondaryly randomly draw n integer, wherein product_num _ifor the number of product i, n is the species number of product; Chromosome as produced is [2,3,5,3, Isosorbide-5-Nitrae, 2,6,5,3,2,3], then represent that the processing batch size of product 1,2,3, L, 12 is followed successively by 2,3,5,3, Isosorbide-5-Nitrae, 2,6,5,3,2,3;

C.2.1.3, Population adaptation angle value calculates

Divide according to buffer size and product batches, the mathematical model that convolution (6), (7), (8), (9), (10), (11) are set up, adopts CPLEX instrument to calculate optimum process-cycle OF _i, after normalization, draw the fitness value NOF of each country _i;

NOF _i＝max _j{OF _j}-OF _i(1)

C.2.1.4, empire divides and selects fitness value NOF _imaximum N _impindividual country is as emperialist, and w in proportion _iall the other country are distributed to corresponding emperialist as colony, forms N _impindividual empire;

$w_i=\left|\frac{{\mathrm{NOF}}_i}{{\mathrm{\Σ}}_{j=1}^{N_{\mathrm{imp}}}{\mathrm{NOF}}_j}\right|---\left(2\right)$

C.2.2, assimilate

Select each colony successively _ijwith the emperialist of its subordinate _i, utilize emperialist _isequence assimilation colony _ijsequence, then C.2.1.3 calculate the colony after assimilation according to step _ijfitness value;

C.2.3, exchange

Contrast each colony successively _ijwith the emperialist of its subordinate _ifitness value, if the former is greater than the latter, then the two exchange status;

C.2.4, the overall fitness value calculation of empire

According to emperialist in each empire _iand colony _ijproduction cycle calculate the overall fitness value NTOF of empire _i;

${\mathrm{TOF}}_i=\mathrm{OF}\left({\mathrm{imperialist}}_i\right)+\mathrm{\ξ}\·\stackrel{\‾}{\mathrm{OF}\left(\mathrm{colony}\right)}---\left(3\right)$

NTOF _i＝max _j{TOF _j}-TOF _i(4)

C.2.5, compete

Choose the colony that in the minimum empire of overall fitness value, fitness value is minimum _ij, calculate each empire and occupy this colonial probability , the N then in stochastic generation one [0,1] scope _impdimension group $R=[r_1,r_2,K,r_{N_{\mathrm{imp}}}],$ Calculating probability array $D=[W_{w_1}-r_1,W_{w_2}-r_2,K,W_{w_{N_{\mathrm{imp}}}}-r_{N_{\mathrm{imp}}}],$ Wherein corresponding probability D _imaximum empire competition triumph, obtains this colony;

$W_{w_i}=\left|\frac{{\mathrm{NTOF}}_i}{{\mathrm{\Σ}}_{j=1}^{N_{\mathrm{imp}}}{\mathrm{NTOF}}_j}\right|---\left(5\right)$

C.2.6, revolution

Select the country that fitness value is minimum, replace original sequence with a probability r random series;

C.2.7, empire is eliminated

But when an empire loses all colony, this empire disappears, and its emperialist is joined in the maximum empire of most fitness value;

C.2.8, kingdom's Competitive Algorithms stops judging

Judge whether the maximum iteration time of kingdom's Competitive Algorithms reaches the fitness value of N or all country identical, if so, then termination algorithm; Otherwise, return step C.2.2;

D, buffer size optimization stop judging

By process-cycle assignment minimum when C.2 middle algorithm stops to T _i, be the process-cycle under current buffer, if T _i=T _i-2, namely the process-cycle continuous three times all constant, then workshop optimize after production capacity be $\mathrm{Throughput}=\frac{\mathrm{Time}\_\mathrm{available}*\mathrm{input}\_\mathrm{num}}{T_{i-2}},$ Time_available is the time span of production prediction, input_num be algorithm input piece count, now buffer size be set to buffer _i-2; Otherwise, return step C.1;

Further, step C.2.1.3 in, described integer programming mathematical model is as follows:

C.2.1.3.1, variable-definition

W: processing station quantity;

C: work-piece batch quantity;

M _i: the quantity of equivalent parallel machine in processing station i;

T: process-cycle;

A _i: the working rate of equipment in processing station i;

B _i: the size of buffer pool size between processing station i and processing station i+1;

N _i: piece count in jth batch;

P _ij: the process time of each workpiece on the equipment of processing station i in jth batch;

D _ij: jth criticizes the total dead time of workpiece at processing station i;

S _ij: jth criticizes the setup time that workpiece is processed at processing station i;

W (i, j, k): processing station i at converted products j-k, and buffer zone i, when 1, L, N-1, is filled the time used by j-k+1, L, j, wherein k=0;

Y (i, j, k): processing station i will be sequentially j-k, j-k+1, L, j, wherein k=0 in the i-1 of buffer zone, the Product processing complete time used of 1, L, N-1;

Integer programming mathematical model is as follows:

C.2.1.3.2, model is set up

min T (6)

$\begin{array}{cc}{d}_{\mathrm{ij}}\≥n_j{P}_{\mathrm{ij}}+s_{\mathrm{ij}}& P_{\mathrm{ij}}=\frac{P_{\mathrm{ij}}}{a_i\·m_i}---\left(7\right)\end{array}$

$T\≥\underset{j=1}{\overset{c}{\mathrm{\Σ}}}{d}_{\mathrm{ij}}---\left(8\right)$

$\underset{r=0}{\overset{k}{\mathrm{\Σ}}}{d}_{i,j-r}\≥\underset{r=0}{\overset{k}{\mathrm{\Σ}}}{d}_{i-1,j-r}-w(i-1,j,k)+P_{\mathrm{ij}}+s_{i,j-k}---\left(9\right)$

$\underset{r=0}{\overset{k}{\mathrm{\Σ}}}{d}_{i,j-r}\≥\underset{r=0}{\overset{k}{\mathrm{\Σ}}}{d}_{i-1,j-r}-Y(i-1,j,k)+P_{i,j-k}+s_{i+1,j-k}---\left(10\right)$

i＝1，2，L，w；j＝1，2，L c；k＝0，1，L，c-1 (11)

In above-mentioned formula, formula (6) is target function type, namely minimizes the production cycle; Formula (7) is the dead time constraint of workpiece on equipment, considers " obstruction " phenomenon that may occur, d _ijat least equal setup time and process time sum; Formula (8) represents production cycle computing formula, is the maximum dead time of processing tasks at all processing stations; Formula (9) represents input work-piece constraint, and consider equipment i and equipment i-1 and buffer zone i, if workpiece collection j-kj-k+1, the size of L, j is less than the size of buffer zone i, then have

$w(i,j,k)=\underset{r=j-k}{\overset{j}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}$ When $\underset{r=j-k}{\overset{j}{\mathrm{\&Sigma;}}}{n}_r<b_i---\left(12\right)$

As workpiece collection j-k, when the size of j-k+1, L, j is greater than the size of buffer zone i, then have and a batch j-k+ η has part γ _ijkload buffer zone i,

${\mathrm{\&gamma;}}_{\mathrm{ijk}}=(b_i-\underset{r=j-k}{\overset{j-k+\mathrm{\&eta;}-1}{\mathrm{\&Sigma;}}}{n}_r)/n_{j-k+\mathrm{\&eta;}}---\left(13\right)$

Now have

$w(i,j,k)=\underset{r=j-k}{\overset{j-k+\mathrm{\&eta;}-1}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}+{\mathrm{\&gamma;}}_{\mathrm{ijk}}(d_{i,j-k+\mathrm{\&eta;}}-{\mathrm{st}}_{i,j-k+\mathrm{\&eta;}})+{\mathrm{st}}_{i,j-k+\mathrm{\&eta;}}---\left(14\right)$

If equipment i-1 completes workpiece collection j-k, the time period of the processing of j-k+1, L, j is [t ₁, t ₂], consider and can fulfil preliminary work ahead of schedule, then the completion date the earliest of equipment i is t ₂+ P _ij, on-stream time is t the latest ₁+ w (i-1, j, k)-s _{j, j-k}, i.e. the hold-up time of machine i be formula (9);

Formula (10) represents output work-piece constraint, if workpiece collection j-k, the size of j-k+1, L, j is less than the size of buffer zone i-1, then have

$Y(i,j,k)=\underset{r=j-k}{\overset{j}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}$ When $\underset{r=j-k}{\overset{j}{\mathrm{\&Sigma;}}}{n}_r<b_{i-1}---\left(15\right)$

As workpiece collection j-k, when the size of j-k+1, L, j is greater than the size of buffer zone i-1, then have and a batch j-η has part τ _ijkbuffer zone i-1 is loaded with workpiece j,

${\mathrm{\&tau;}}_{\mathrm{ijk}}=(b_{i-1}-\underset{r=j-\mathrm{\&eta;}}{\overset{j}{\mathrm{\&Sigma;}}}{n}_r)/n_{j-\mathrm{\&eta;}}---\left(16\right)$

Now have

$Y(i,j,k)=\underset{r=j-\mathrm{\&eta;}}{\overset{j}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}+{\mathrm{\&tau;}}_{\mathrm{ijk}}(d_{i,j+\mathrm{\&eta;}}-{\mathrm{st}}_{i,j+\mathrm{\&eta;}})+{\mathrm{st}}_{i,j+\mathrm{\&eta;}}---\left(17\right)$

Consideration equipment i and equipment i+1 and buffer zone i+1, if equipment i+1 completes workpiece collection j-kj-k+1, the time period of the processing of L, j is [t ₃, t ₄], consider and can fulfil preliminary work ahead of schedule, then the completion date the earliest of equipment i is t4-Y (i+1, j, k), and on-stream time is t ₃+ si _{i+1, j-k}-p _{i, j-k}, i.e. the hold-up time of machine i $\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i,j-r}\&GreaterEqual;t_4-Y(i+1,j,k)-(t_3+{\mathrm{st}}_{i+1,j-k}-p_{i,j-k}),$ Be formula (10);

Formula (11) is Integer constrained characteristic;

E, production line Productivity Allocation show with control result; User is permitted by module, and above optimal control result is sent to server, and workshop management personnel execute the task according to result.

The present invention has the following advantages compared with prior art:

1, the present invention has considered " obstruction " and " hunger " phenomenon in activity in production, and establishes Productivity Allocation method, realizes the optimal sequencing of work-piece batch, has heightened the efficiency of Productivity Allocation.

2, the present invention adopts batch division of kingdom's Competitive Algorithms to different product to be optimized, and avoids the limitation that traditional production capacity optimized algorithm only considers bottleneck station, achieves the optimization of overall importance of production capacity.

3, present invention further contemplates the adjustability of equipment room buffer size and the optimization of product batches processing sequence, eliminating in activity in production while " obstruction " and " hunger " phenomenon as far as possible, further the production capacity of production line being optimized.

Accompanying drawing explanation

Below in conjunction with the drawings and specific embodiments, the present invention is further detailed explanation.

Fig. 1 is process flow diagram of the present invention.

Fig. 2 is control system block diagram of the present invention.

Embodiment

As shown in Figure 1 and Figure 2, performing step of the present invention is as follows:

Motor product of production line information, by authentication module License Info, is input to control system by browser by A, user;

B, information pre-processing.Transform that production data is concrete Productivity Allocation model (see C.2.1.3.1 and C.2.1.3.), user arranges production line model parameter, control strategy parameter and control system interaction protocol.

C, production line Productivity Allocation and control module; Specifically comprise production line buffer size regulator module, kingdom's Competitive Algorithms optimizing product batch divides submodule and product batches production sequence regulator module.

C.1, production buffer district size control submodule is arranged

Equipment room buffer size initial value buffer is set ₀with the incremental change Δ of each circulation,

buffer _i＝buffer ₀+i*Δ

C.2, kingdom's Competitive Algorithms search (kingdom's Competitive Algorithms optimizing product batch divides submodule, calculates the production cycle);

The buffer size of equipment room calculated according to step 1 and the processing batch size of often kind of product after optimizing, set up integer programming model, calculate the optimum process-cycle, as the basis of each kingdom Competitive Algorithms ideal adaptation angle value, process-cycle after the Competitive Algorithms convergence of final kingdom, batch size corresponding with it and product batches processing sequence were as production capacity prioritization scheme as the rreturn value of step C.

C.2.1, kingdom's Competitive Algorithms initialization

C.2.1.1, algorithm parameter arrange, setting parameter, the country number N of set algorithm _pop, emperialist number N _imp, colony weight factor ξ, revolution rate r and maximum cycle N;

C.2.1.2, initial population is produced

Counter i=0 is set, at [1, product_num _i] interior repetition N _popsecondaryly randomly draw n integer, wherein product_num _ifor the number of product i, n is the species number of product; Chromosome as produced is [2,3,5,3, Isosorbide-5-Nitrae, 2,6,5,3,2,3], then represent that the processing batch size of product 1,2,3, L, 12 is followed successively by 2,3,5,3, Isosorbide-5-Nitrae, 2,6,5,3,2,3;

C.2.1.3, Population adaptation angle value calculates

Divide according to buffer size and product batches, the mathematical model that convolution (6), (7), (8), (9), (10), (11) are set up, adopts CPLEX instrument to calculate optimum process-cycle OF _i, after normalization, draw the fitness value NOF of each country _i;

NOF _i＝max _j{OF _j}-OF _i(1)

C.2.1.4, empire divides and selects fitness value NOF _imaximum N _impindividual country is as emperialist, and w in proportion _iall the other country are distributed to corresponding emperialist as colony, forms N _impindividual empire;

$w_i=\left|\frac{{\mathrm{NOF}}_i}{{\mathrm{\&Sigma;}}_{j=1}^{N_{\mathrm{imp}}}{\mathrm{NOF}}_j}\right|---\left(2\right)$

C.2.2, assimilate

Select each colony successively _ijwith the emperialist of its subordinate _i, utilize emperialist _isequence assimilation colony _ijsequence, then C.2.1.3 calculate the colony after assimilation according to step _ijfitness value;

C.2.3, exchange

Contrast each colony successively _ijwith the emperialist of its subordinate _ifitness value, if the former is greater than the latter, then the two exchange status;

C.2.4, the overall fitness value calculation of empire

According to emperialist in each emptre _iand colony _ijproduction cycle calculate the overall fitness value NTOF of empire _i;

${\mathrm{TOF}}_i=\mathrm{OF}\left({\mathrm{imperialist}}_i\right)+\mathrm{\&xi;}\&CenterDot;\stackrel{\&OverBar;}{\mathrm{OF}\left(\mathrm{colony}\right)}---\left(3\right)$

NTOF _i＝max _j{TOF _j}-TOF _i(4)

C.2.5, compete

Choose the colony that in the minimum empire of overall fitness value, fitness value is minimum _ij, calculate each empire and occupy this colonial probability then the N in stochastic generation one [0,1] scope _impdimension group $R=[r_1,r_2,K,r_{N_{\mathrm{imp}}}],$ Calculating probability array $D=[W_{w_1}-r_1,W_{w_2}-r_2,K,W_{w_{N_{\mathrm{imp}}}}-r_{N_{\mathrm{imp}}}],$ Wherein corresponding probability D _imaximum empire competition triumph, obtains this colony.

$W_{w_i}=\left|\frac{{\mathrm{NTOF}}_i}{{\mathrm{\&Sigma;}}_{j=1}^{N_{\mathrm{imp}}}{\mathrm{NTOF}}_j}\right|---\left(5\right)$

C.2.6, revolution

Select the country that fitness value is minimum, replace original sequence with a probability r random series;

C.2.7, empire is eliminated

But when an empire loses all colony, this empire disappears, and its emperialist is joined in the maximum empire of most fitness value;

C.2.8, kingdom's Competitive Algorithms stops judging

Judge whether the maximum iteration time of kingdom's Competitive Algorithms reaches the fitness value of N or all country identical, if so, then termination algorithm; Otherwise, return step C.2.2;

D, buffer size optimization stop judging

By process-cycle assignment minimum when C.2 middle algorithm stops to T _i, be the process-cycle under current buffer, if T _i=T _i-2, namely the process-cycle continuous three times all constant, then workshop optimize after production capacity be $\mathrm{Throughput}=\frac{\mathrm{Time}\_\mathrm{available}*\mathrm{input}\_\mathrm{num}}{T_{i-2}}$ (Time_available is the time span of production prediction, input_num be algorithm input piece count), now buffer size be set to buffer _i-2; Otherwise, return step C.1;

Further, step C.2.1.3 in, described integer programming mathematical model is as follows:

C.2.1.3.1, variable-definition

W: processing station quantity;

C: work-piece batch quantity;

M _i: the quantity of equivalent parallel machine in processing station i;

T: process-cycle;

A _i: the working rate of equipment in processing station i;

B _i: the size of buffer pool size between processing station i and processing station i+1;

N _i: piece count in jth batch;

P _ij: the process time of each workpiece on the equipment of processing station i in jth batch;

D _ij: jth criticizes the total dead time of workpiece at processing station i;

S _ij: jth criticizes the setup time that workpiece is processed at processing station i;

W (i, j, k): processing station i at converted products j-k, and buffer zone i, when 1, L, N-1, is filled the time used by j-k+1, L, j, wherein k=0;

Y (i, j, k): processing station i will be sequentially j-k, j-k+1, L, j, wherein k=0 in the i-1 of buffer zone, the Product processing complete time used of 1, L, N-1;

Integer programming mathematical model is as follows:

C.2.1.3.2, model is set up

min T (6)

$\begin{array}{cc}{d}_{\mathrm{ij}}\&GreaterEqual;n_j{P}_{\mathrm{ij}}+s_{\mathrm{ij}}& P_{\mathrm{ij}}=\frac{P_{\mathrm{ij}}}{a_i\&CenterDot;m_i}---\left(7\right)\end{array}$

$T\&GreaterEqual;\underset{j=1}{\overset{c}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ij}}---\left(8\right)$

$\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i,j-r}\&GreaterEqual;\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i-1,j-r}-w(i-1,j,k)+P_{\mathrm{ij}}+s_{i,j-k}---\left(9\right)$

$\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i,j-r}\&GreaterEqual;\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i-1,j-r}-Y(i-1,j,k)+P_{i,j-k}+s_{i+1,j-k}---\left(10\right)$

i＝1，2，L，w；j＝1，2，L c；k＝0，1，L，c-1 (11)

In above-mentioned formula, formula (6) is target function type, namely minimizes the production cycle; Formula (7) is the dead time constraint of workpiece on equipment, considers " obstruction " phenomenon that may occur, d _ijat least equal setup time and process time sum; Formula (8) represents production cycle computing formula, is the maximum dead time of processing tasks in all processing stations (comprising multiple stage type of the same race); Formula (9) represents input work-piece constraint, and consider equipment i and equipment i-1 and buffer zone i, if workpiece collection j-k, the size of j-k+1, L, j is less than the size of buffer zone i, then have

$w(i,j,k)=\underset{r=j-k}{\overset{j}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}$ When $\underset{r=j-k}{\overset{j}{\mathrm{\&Sigma;}}}{n}_r<b_i---\left(12\right)$

As workpiece collection j-k, when the size of j-k+1, L, j is greater than the size of buffer zone i, then have and a batch j-k+ η has part γ _ijkload buffer zone i,

${\mathrm{\&gamma;}}_{\mathrm{ijk}}=(b_i-\underset{r=j-k}{\overset{j-k+\mathrm{\&eta;}-1}{\mathrm{\&Sigma;}}}{n}_r)/n_{j-k+\mathrm{\&eta;}}---\left(13\right)$

Now have

$w(i,j,k)=\underset{r=j-k}{\overset{j-k+\mathrm{\&eta;}-1}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}+{\mathrm{\&gamma;}}_{\mathrm{ijk}}(d_{i,j-k+\mathrm{\&eta;}}-{\mathrm{st}}_{i,j-k+\mathrm{\&eta;}})+{\mathrm{st}}_{i,j-k+\mathrm{\&eta;}}---\left(14\right)$

If equipment i-1 completes workpiece collection j-k, the time period of the processing of j-k+1, L, j is [t ₁, i ₂], consider and can fulfil preliminary work ahead of schedule, then the completion date the earliest of equipment i is t ₂+ P _ij, on-stream time is t the latest ₁+ w (i-1, j, k)-s _{j, j-k}, i.e. the hold-up time of machine i be formula (9);

Formula (10) represents output work-piece constraint, if workpiece collection j-k, the size of j-k+1, L, j is less than the size of buffer zone i-1, then have

$Y(i,j,k)=\underset{r=j-k}{\overset{j}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}$ When $\underset{r=j-k}{\overset{j}{\mathrm{\&Sigma;}}}{n}_r<b_{i-1}---\left(15\right)$

As workpiece collection j-k, when the size of j-k+1, L, j is greater than the size of buffer zone i-1, then have and a batch j-η has part τ _ijkbuffer zone i-1 is loaded with workpiece j,

${\mathrm{\&tau;}}_{\mathrm{ijk}}=(b_{i-1}-\underset{r=j-\mathrm{\&eta;}}{\overset{j}{\mathrm{\&Sigma;}}}{n}_r)/n_{j-\mathrm{\&eta;}}---\left(16\right)$

Now have

$Y(i,j,k)=\underset{r=j-\mathrm{\&eta;}}{\overset{j}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}+{\mathrm{\&tau;}}_{\mathrm{ijk}}(d_{i,j+\mathrm{\&eta;}}-{\mathrm{st}}_{i,j+\mathrm{\&eta;}})+{\mathrm{st}}_{i,j+\mathrm{\&eta;}}---\left(17\right)$

Consideration equipment i and equipment i+1 and buffer zone i+1, if equipment i+1 completes workpiece collection j-kj-k+1, the time period of the processing of L, j is [t ₃, t ₄], consider and can fulfil preliminary work ahead of schedule, then the completion date the earliest of equipment i is t ₄-Y (i+1, j, k), on-stream time is t ₃+ st _{t+1, j-k}- _{i, j-k}, i.e. the hold-up time of machine i $\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i,j-r}\&GreaterEqual;t_4-Y(i+1,j,k)-(t_3+{\mathrm{st}}_{i+1,j-k}-p_{i,j-k}),$ Be formula (10);

Formula (11) is Integer constrained characteristic;

E, production line Productivity Allocation show with control result; User is permitted by module, and above optimal control result is sent to server, and workshop management personnel execute the task according to result.

The object of the invention is to solve existing mix flow Productivity Allocation and control method poor for applicability, the problem that Optimal performance is not high, provide a kind of mix flow Productivity Allocation method based on kingdom's Competitive Algorithms and control system, to facilitate Productivity Allocation and the adjustment activity of enterprise, and improve production efficiency and the production chains of mix flow.

Claims (1)

1. a mix flow Productivity Allocation method, is characterized in that: performing step is as follows:

Motor product of production line information, by authentication module License Info, is input to control system by browser by A, user;

B, information pre-processing; Transform that production data is concrete Productivity Allocation model, user arranges production line model parameter, control strategy parameter and control system interaction protocol;

C, production line Productivity Allocation and control module; Specifically comprise production line buffer size regulator module, kingdom's Competitive Algorithms optimizing product batch divides submodule and product batches production sequence regulator module;

C.1, production buffer district size control submodule is arranged

Equipment room buffer size initial value buffer is set ₀with the incremental change Δ of each circulation;

buffer _i＝buffer ₀+i*Δ

C.2, kingdom's Competitive Algorithms search;

The buffer size of equipment room calculated according to step 1 and the processing batch size of often kind of product after optimizing, set up integer programming model, calculate the optimum process-cycle, as the basis of each kingdom Competitive Algorithms ideal adaptation angle value, process-cycle after the Competitive Algorithms convergence of final kingdom, batch size corresponding with it and product batches processing sequence were as production capacity prioritization scheme as the rreturn value of step C;

C.2.1, kingdom's Competitive Algorithms initialization

C.2.1.1, algorithm parameter arrange, setting parameter, the country number N of set algorithm _pop, emperialist number N _imp, colony weight factor ξ, revolution rate r and maximum cycle N;

C.2.1.2, initial population is produced

Counter i=0 is set, at [1, product_num _i] interior repetition N _popsecondaryly randomly draw n integer, wherein product_num _ifor the number of product i, n is the species number of product; Chromosome as produced is [2,3,5,3, Isosorbide-5-Nitrae, 2,6,5,3,2,3], then represent that the processing batch size of product 1,2,3, L, 12 is followed successively by 2,3,5,3, Isosorbide-5-Nitrae, 2,6,5,3,2,3;

C.2.1.3, Population adaptation angle value calculates

Divide according to buffer size and product batches, the mathematical model that convolution (6), (7), (8), (9), (10), (11) are set up, adopts CPLEX instrument to calculate optimum process-cycle OF _i, after normalization, draw the fitness value NOF of each country _i;

NOF _i＝max _j{OF _j}-OF _i(1)

C.2.1.4, empire divides and selects fitness value NOF _imaximum N _impindividual country is as emperialist, and w in proportion _iall the other country are distributed to corresponding emperialist as colony, forms N _impindividual empire;

$w_i=\left|\frac{{\mathrm{NOF}}_i}{{\mathrm{\&Sigma;}}_{j=1}^{N_{\mathrm{imp}}}{\mathrm{NOF}}_j}\right|---\left(2\right)$

C.2.2, assimilate

Select each colony successively _ijwith the emperialist of its subordinate _i, utilize emperialist _isequence assimilation colony _ijsequence, then C.2.1.3 calculate the colony after assimilation according to step _ijfitness value;

C.2.3, exchange

Contrast each colony successively _ijwith the emperialist of its subordinate _ifitness value, if the former is greater than the latter, then the two exchange status;

C.2.4, the overall fitness value calculation of empire

According to emperialist in each empire _iand colony _ijproduction cycle calculate the overall fitness value NTOF of empire _j;

${\mathrm{TOF}}_i=\mathrm{OF}\left({\mathrm{imperialist}}_i\right)+\mathrm{\&xi;}\&CenterDot;\stackrel{\&OverBar;}{\mathrm{OF}\left(\mathrm{colony}\right)}---\left(3\right)$

NTOF _i＝max _j{TOF _j}-TOF _i(4)

C.2.5, compete

Choose the colony that in the minimum empire of overall fitness value, fitness value is minimum _ij, calculate each empire and occupy this colonial probability then the N in stochastic generation one [0,1] scope _impdimension group $R=[r_1,r_2,K,r_{N_{\mathrm{imp}}}],$ Calculating probability array $D=[W_{w_1}-r_1,W_{w_2}-r_2,K,W_{w_{N_{\mathrm{imp}}}}-r_{N_{\mathrm{imp}}}],$ Wherein corresponding probability D _imaximum empire competition triumph, obtains this colony;

$W_{w_i}=\left|\frac{{\mathrm{NTOF}}_i}{{\mathrm{\&Sigma;}}_{j=1}^{N_{\mathrm{imp}}}{\mathrm{NTOF}}_j}\right|---\left(5\right)$

C.2.6, revolution

Select the country that fitness value is minimum, replace original sequence with a probability r random series;

C.2.7, empire is eliminated

But when an empire loses all colony, this empire disappears, and its emperialist is joined in the maximum empire of most fitness value;

C.2.8, kingdom's Competitive Algorithms stops judging

Judge whether the maximum iteration time of kingdom's Competitive Algorithms reaches the fitness value of N or all country identical, if so, then termination algorithm; Otherwise, return step C.2.2;

D, buffer size optimization stop judging

By process-cycle assignment minimum when C.2 middle algorithm stops to T _i, be the process-cycle under current buffer, if T _i=T _i-2, namely the process-cycle continuous three times all constant, then workshop optimize after production capacity be $\mathrm{Throughput}=\frac{\mathrm{Time}\_\mathrm{available}*\mathrm{input}\_\mathrm{num}}{T_{i-2}},$ Time_available is the time span of production prediction, input_num be algorithm input piece count, now buffer size be set to buffer _i-2; Otherwise, return step C.1;

Further, step C.2.1.3 in, described integer programming mathematical model is as follows:

C.2.1.3.1, variable-definition

W: processing station quantity;

C: work-piece batch quantity;

M _i: the quantity of equivalent parallel machine in processing station i;

T: process-cycle;

A _i: the working rate of equipment in processing station i;

B _i: the size of buffer pool size between processing station i and processing station i+1;

N _i: piece count in jth batch;

P _ij: the process time of each workpiece on the equipment of processing station i in jth batch;

D _ij: jth criticizes the total dead time of workpiece at processing station i;

S _ij: jth criticizes the setup time that workpiece is processed at processing station i;

W (i, j, k): processing station i at converted products j-k, and buffer zone i, when 1, L, N-1, is filled the time used by j-k+1, L, j, wherein k=0;

Y (i, j, k): processing station i will be sequentially j-k, j-k+1, L, j, wherein k=0 in the i-1 of buffer zone, the Product processing complete time used of 1, L, N-1;

Integer programming mathematical model is as follows:

C.2.1.3.2, model is set up

min T (6)

$\begin{array}{cc}{d}_{\mathrm{ij}}\&GreaterEqual;n_j{P}_{\mathrm{ij}}+s_{\mathrm{ij}}& P_{\mathrm{ij}}=\frac{p_{\mathrm{ij}}}{a_i\&CenterDot;m_i}\end{array}---\left(7\right)$

$T\&GreaterEqual;\underset{j=1}{\overset{c}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ij}}---\left(8\right)$

$\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i,j-r}\&GreaterEqual;\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i-1,j-r}-w(i-1,j,k)+P_{\mathrm{ij}}+s_{i,j-k}---\left(9\right)$

$\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i,j-r}\&GreaterEqual;\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i+1,j-r}-Y(i+1,j,k)+P_{i,j-k}+s_{i+1,j-k}---\left(10\right)$

i＝1，2，L，w；j＝1，2，L c；k＝0，1，L，c-1 (11)

In above-mentioned formula, formula (6) is target function type, namely minimizes the production cycle; Formula (7) is the dead time constraint of workpiece on equipment, considers " obstruction " phenomenon that may occur, d _ijat least equal setup time and process time sum; Formula (8) represents production cycle computing formula, is the maximum dead time of processing tasks at all processing stations; Formula (9) represents input work-piece constraint, and consider equipment i and equipment i-1 and buffer zone i, if workpiece collection j-k, the size of j-k+1, L, j is less than the size of buffer zone i, then have

$w(i,j,k)=\underset{r=j-k}{\overset{j}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}$ When $\underset{r=j-k}{\overset{j}{\mathrm{\&Sigma;}}}{n}_r<b_i---\left(12\right)$

As workpiece collection j-k, when the size of j-k+1, L, j is greater than the size of buffer zone i, then have and a batch j-k+ η has part γ _ijkload buffer zone i,

${\mathrm{\&gamma;}}_{\mathrm{ijk}}=(b_i-\underset{r=j-k}{\overset{j-k+\mathrm{\&eta;}-1}{\mathrm{\&Sigma;}}}{n}_r)/n_{j-k+\mathrm{\&eta;}}---\left(13\right)$

Now have

$w(i,j,k)=\underset{r=j-k}{\overset{j-k+\mathrm{\&eta;}-1}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}+{\mathrm{\&gamma;}}_{\mathrm{ijk}}(d_{i,j-k+\mathrm{\&eta;}}-{\mathrm{st}}_{i,j-k+\mathrm{\&eta;}})+{\mathrm{st}}_{i,j-k+\mathrm{\&eta;}}---\left(14\right)$

If equipment i-1 completes workpiece collection j-k, the time period of the processing of j-k+1, L, j is [t ₁, t ₂], consider and can fulfil preliminary work ahead of schedule, then the completion date the earliest of equipment i is t ₂+ P _ij, on-stream time is t the latest _l+ w (i-1, j, k)-s _{j, j-k}, i.e. the hold-up time of machine i $\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i,j-r}\&GreaterEqual;t_2+P_{\mathrm{ij}}-(t_1+w(i-1,j,k)-s_{j,j-k}),$ Be formula (9);

Formula (10) represents output work-piece constraint, if workpiece collection j-k, the size of j-k+1, L, j is less than the size of buffer zone i-1, then have

$Y(i,j,k)=\underset{r=j-j}{\overset{j}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}$ When $\underset{r=j-k}{\overset{j}{\mathrm{\&Sigma;}}}{n}_r<b_{i-1}---\left(15\right)$

As workpiece collection j-k, when the size of j-k+1, L, j is greater than the size of buffer zone i-1, then have and a batch j-η has part τ _ijkbuffer zone i-1 is loaded with workpiece j,

${\mathrm{\&tau;}}_{\mathrm{ijk}}=(b_{i-1}-\underset{r=j-\mathrm{\&eta;}}{\overset{j}{\mathrm{\&Sigma;}}}{n}_r)/n_{j-\mathrm{\&eta;}}---\left(16\right)$

Now have

$Y(i,j,k)=\underset{r=j-\mathrm{\&eta;}}{\overset{j}{\mathrm{\&Sigma;}}}{d}_{\mathrm{ir}}+{\mathrm{\&tau;}}_{\mathrm{ijk}}(d_{i,j+\mathrm{\&eta;}}-{\mathrm{st}}_{i,j+\mathrm{\&eta;}})+{\mathrm{st}}_{i,j+\mathrm{\&eta;}}---\left(17\right)$

Consideration equipment i and equipment i+1 and buffer zone i+1, if equipment i+1 completes workpiece collection j-k, the time period of the processing of j-k+1, L, j is [t ₃, t ₄], consider and can fulfil preliminary work ahead of schedule, then the completion date the earliest of equipment i is t ₄-Y (i+1, j, k), on-stream time is t ₃+ st _{i+1, j-k}-p _{i, j-k}, i.e. the hold-up time of machine i $\underset{r=0}{\overset{k}{\mathrm{\&Sigma;}}}{d}_{i,j-r}\&GreaterEqual;t_4-Y(i+1,j,k)-(t_3+{\mathrm{st}}_{i+1,j-k}-p_{i,j-k}),$ Be formula (10);

Formula (11) is Integer constrained characteristic;

E, production line Productivity Allocation show with control result; User is permitted by module, and above optimal control result is sent to server, and workshop management personnel execute the task according to result.