CN102768737B - A kind of job shop bottleneck identification method considering multidimensional feature attribute of machine - Google Patents

Abstract

The present invention proposes a kind of job shop bottleneck identification method considering multidimensional feature attribute of machine.First optimized algorithm is adopted to carry out job scheduling Optimization Solution, secondly the way of tradition according to machine one-sided factor identification bottleneck is changed, consider many-sided characteristic attribute of bottleneck, propose many attributes bottleneck identification method, comprehensive thoroughly evaluating machine performance carries out bottleneck identification.The present invention by bottleneck utilize and bottleneck identification problem be placed on Unified frame under carry out integrated solving, solve traditional bottleneck identification cut off mutually with prioritization scheme and cause bottleneck identification forbidden, the not excellent deficiency of prioritization scheme; After obtaining optimizing scheduling scheme, consider the different characteristic attribute of machine, adopt TOPSIS many attributes bottleneck identification method to carry out bottleneck identification, overcome the defect that traditional bottleneck distinguishing indexes is unilateral, recognition result is biased.

Description

A kind of job shop bottleneck identification method considering multidimensional feature attribute of machine

Technical field

The present invention relates to job shop bottleneck identification technical field, be specially a kind of job shop bottleneck identification method considering multidimensional feature attribute of machine.

Background technology

The finiteness of manufacturing recourses and production system itself have statistical fluctuations and processing dependence, and restriction system throughput must be caused to maximize " bottleneck " phenomenon exported.Bounding theory (Theory of Constraints, TOC) think that bottleneck (Bottleneck) is the reference mark really restricting whole system throughput and inventory level, loss on bottleneck means the loss of whole production system, only base oneself upon bottleneck and bottleneck utilization factor is maximized, entire system output just can be made optimum.Therefore, the dependence point of production planning and control and basic point should be placed on bottleneck.

Existing bottleneck identification research concentrates on bottleneck identification index and bottleneck identification method, and existing bottleneck identification index is divided in goods class (Work-in-process Orientation), capacity of equipment class (Equipment Capacity Orientation) two classes; Bottleneck identification method is divided into index Direct Recognition method, mathematical methods and data analysis method three class.Bottleneck identification method is mutually corresponding with bottleneck identification index, is identified by bottleneck index:

(1) index Direct Recognition method, directly carries out bottleneck identification by production scene machining situation or in goods index.The bottleneck defined in document comprises: the machine with the longest average latency; There is the machine of maximum queue length; The machine that system working ability is the poorest; The machine that load is maximum; The machine that overall equipment efficiency.comprehensive efficiency of equipment (Overall EquipmentEffectiveness, OEE) is maximum.

(2) mathematical methods, in the performance parameter of hypothesis machine, as failure rate, process-cycle, MTTR etc., all meet on the basis of certain probability distribution, set up the mathematical model of production line, by the output of analytic engine on the bottleneck machine affecting recognition system of system produce.

(3) data analysis method, monitors in real time based on emulation or online data, carries out treatment and analysis, the bottleneck machine of discrimination system for emulated data, real time data etc.

Traditional work workshop bottleneck identification considers one-sided factor, considers multifactor few.Most document considers bottleneck in a certain respect, set up bottleneck identification index based on single factor test feature and carry out bottleneck identification, multiple index considered by minority document, but is all based on simple rule, the quantity of information of machine characteristic attribute can not be excavated, the object of comprehensive evaluation machine cannot be realized.

Summary of the invention

The technical matters solved

For solving prior art Problems existing, the present invention proposes a kind of job shop bottleneck identification method considering multidimensional feature attribute of machine.

Technical scheme

The present invention uses the novel optimization operation logic of TOC, integrated solving is carried out under proposing that bottleneck utilization and bottleneck identification problem are placed on Unified frame, first level is that bottleneck utilizes, adopt optimized algorithm to carry out job scheduling Optimization Solution, ensure that job scheduling scheme overall performance optimum and bottleneck machine capability make full use of degree the highest; Second level is bottleneck identification, and change the way of tradition according to machine one-sided factor identification bottleneck, consider many-sided characteristic attribute of bottleneck, propose many attributes bottleneck identification method, comprehensive thoroughly evaluating machine performance carries out bottleneck identification.This framework changes the way of traditional bottleneck identification independent of optimizing scheduling scheme, solving job shop scheduling problem optimization is combined with bottleneck identification, first carry out bottleneck and make full use of the identification carrying out system bottleneck again, both ensure that making full use of of bottleneck device, in turn ensure that the total optimization of optimizing scheduling scheme.

The bottleneck of indication of the present invention utilizes the core missions of layer to be the Optimization Solution carrying out Job Shop scheduling problem.

The core missions of the bottleneck identification layer of indication of the present invention are identification bottleneck machines.

The bottleneck that the present invention identifies is the maximum machine of the evaluation of estimate of being tried to achieve by multiple attributive decision making method.

Technical scheme of the present invention is:

A kind of described job shop bottleneck identification method considering multidimensional feature attribute of machine, is characterized in that: comprise the following steps:

Step 1: the optimal scheduling scheme Ω adopting optimized algorithm determination job shop:

Ω＝{{B ₁₁,…,B _v1,…,B _e1;C ₁₁,…,C _vl,…,C _e1;Z _l1,…,Z _v1,…,Z _e1},…,{B _li,…,B _vi,…,B _ei;C _li,…,C _vi,…,C _ei;Z _1i,…,Z _vi,…,Z _ei},…,{B _1m,…,B _vm,…,B _em;C _1m,…,C _vm,…,C _em;Z _1m,…,Z _vm,…,Z _em}}

Wherein, B _virepresent v ∈ E={1,2 ..., e} workpiece at the i-th ∈ M={1,2 ..., the beginning process time on m} platform machine, C _virepresent the machine time of v workpiece on i-th machine, Z _virepresent the frock setup time of v workpiece on i-th machine;

Step 2: set up machine and integrate as A={A ₁, A ₂..., A _m, evaluation attributes integrate as X={X ₁, X ₂..., X _n; According to the optimal scheduling scheme Ω that step 1 obtains, calculate each evaluation attributes value of each machine, obtain the decision matrix D=(x for bottleneck identification _ij) _{m × n}, x _ijit is a jth evaluation attributes value of the machine of i-th;

Step 3: adopt vectorial standardized method by the decision matrix D=(x of step 2 _ij) _{m × n}convert criteria decision matrix R=(r to _ij) _{m × n}:

$r_{\mathrm{ij}}=x_{\mathrm{ij}}/\sqrt{\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{x}_{\mathrm{ij}}^2}$

Step 4: adopt entropy assessment determination evaluation attributes weight, evaluation attributes weight vectors W=(ω ₁, ω ₂..., ω _j..., ω _n):

Step 4.1: by the standardization of each evaluation attributes value, wherein the nondimensionalization property value p of lower i-th machine of a jth attribute _ijfor:

$p_{\mathrm{ij}}=x_{\mathrm{ij}}/\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{x}_{\mathrm{ij}}$

Step 4.2: the entropy E calculating a jth evaluation attributes _j:

$E_j=-\mathrm{\δ}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{p}_{\mathrm{ij}}\ln p_{\mathrm{ij}}$

Wherein δ=1/lnm, ensure that 0≤E _j≤ 1;

Step 4.3: the errored message degree d calculating a jth attribute _j:

d _j＝1-E _j

Step 4.4: calculating weight is normalized to errored message degree:

When decision maker does not have the preference between evaluation attributes:

${\mathrm{\ω}}_j=d_j/\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{d}_j$

When decision maker has the preference between evaluation attributes:

${\mathrm{\ω}}_j={\mathrm{\λ}}_j{d}_j/\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\λ}}_j{d}_j$

Wherein λ _jfor the subjective weight of decision maker;

Step 5: the evaluation attributes weight inclusive criteria decision matrix R=(r that step 4 is obtained _ij) _{m × n}in, obtain weighting standard decision matrix V=(v _ij) _{m × n}=(ω _jr _ij) _{m × n};

Step 6: the positive and negative ideal solution A determining evaluation attributes ⁺and A ^-:

$A^{+}=\{\left(\underset{i}{\max }{v}_{\mathrm{ij}}\right|j\∈J),\left(\underset{i}{\min }{v}_{\mathrm{ij}}\right|j\∈J^{\′})|i=M\}=\{v_1^{+},v_2^{+},...,v_j^{+},...,v_n^{+}\}$

$A^{-}=\{\left(\underset{i}{\min }{v}_{\mathrm{ij}}\right|j\∈J),\left(\underset{i}{\max }{v}_{\mathrm{ij}}\right|j\∈J^{\′})|i\∈M\}=\{v_1^{-},v_2^{-},...,v_j^{-},...,v_n^{-}\}$

Wherein J is the set of profit evaluation model evaluation attributes, and J ' is the set of cost type evaluation attributes;

Step 7: tie up Euclid apart from the distance between each machine of calculating and positive ideal solution and minus ideal result by n: every platform machine to the distance of positive ideal solution is:

$S_{i^{+}}=\sqrt{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{(v_{\mathrm{ij}}-v_j^{+})}^2}$

Every platform machine to the distance of minus ideal result is:

$S_{i^{-}}=\sqrt{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{(v_{\mathrm{ij}}-v_j^{-})}^2}$

Step 8: calculate every platform machine and positive ideal solution A ⁺approach degree C _i:

$C_i=S_{i^{-}}/(S_{i^{+}}+S_{i^{-}}),0<C_i<1$

Obtain approach degree C _imaximum machine is the bottleneck of job shop.

A kind of described job shop bottleneck identification method considering multidimensional feature attribute of machine, is characterized in that: evaluation attributes comprise machine burden, machinery utilization rate, active time, enlivens deviation, susceptibility, unit throughput, in progress queues length and overall equipment efficiency.comprehensive efficiency of equipment.

Beneficial effect

The present invention considers the multiple characteristic attribute of machine and carries out bottleneck identification, set up many attributes bottleneck identification model, give the mathematicization definition of bottleneck, bottleneck identification and bottleneck is adopted to utilize Unified frame, bottleneck utilizes layer to be optimized scheduling, bottleneck identification layer based on TOPSIS multiple attributive decision making method, comprehensive evaluation machine, thus identify job shop bottleneck machine, its advantage is:

(1) integrated solving is carried out under bottleneck utilization and bottleneck identification problem being placed on Unified frame, solve traditional bottleneck identification cut off mutually with prioritization scheme and cause that bottleneck identification is inaccurate, the not excellent deficiency of prioritization scheme, while generation optimizing scheduling scheme, determine corresponding bottleneck, the optimal scheduling scheme obtained has more directive significance.

(2) after obtaining optimizing scheduling scheme, consider the different characteristic attribute of machine, adopt TOPSIS many attributes bottleneck identification method to carry out bottleneck identification, overcome the defect that traditional bottleneck distinguishing indexes is unilateral, recognition result is biased.In addition, introduce entropy assessment and make the weight method more science determining machine assessment attribute.

(3) multiple attribute decision making (MADM) of bottleneck identification is except providing system bottleneck, gives judging quota---the approach degree C that prediction becomes bottleneck possibility _i.The priority of possibility bottleneck is become, for the operation of protection bottleneck, secondary bottleneck ability and prevention bottleneck shifting provide important information according to the measurable machine of approach degree index.

Accompanying drawing explanation

Fig. 1: bottleneck identification process sketch of the present invention;

Fig. 2: bottleneck identification procedure chart of the present invention;

Embodiment

The present invention may be used in job shop production control process.Identify bottleneck by method of the present invention, make dispatcher rationally pay close attention to bottleneck and non-bottleneck, improve the efficiency of organization of production, the production capacity in workshop is maximized the use, increases economic efficiency.

Select 24 standard examples of JSSP problem LA class as compare test example in the present embodiment, be respectively the typical problem of 6 kinds of different scales of 10 × 5,15 × 5,10 × 10,15 × 10,20 × 10,30 × 10, often kind of scale chooses 4 examples, totally 24 examples.And method of the present invention and moving bottleneck method of identification (Shifting bottleneckdetection method, SBD) and orthogonal test bottleneck identification method are compared.

The job shop bottleneck identification method of the consideration multidimensional feature attribute of machine in the present embodiment, comprises the following steps:

Step 1: adopt Immune Evolutionary Algorithm CHIEA (Clonal Selection and Hyper Mutations BasedImmune Evolution Algorithm) to determine the optimal scheduling scheme Ω of job shop, target is the ability making full use of bottleneck, realizes the optimum of scheduling scheme overall performance index.Immune Evolutionary Algorithm is shown in document " Immune Evolutionary Algorithm solves static Job shop and dispatches " (Niu Ganggang, Sun Shudong, remaining for army building, mechanical engineering journal, 2006,42 (05): 87-91.):

It is as follows that Immune Evolutionary Algorithm CHIEA solves optimal scheduling protocol step:

Step a constructs antigen and initial antibodies.Have the antigen based on coding of machine preference according to input job shop scheduling problem and related constraint information stochastic generation, this coded system compares tradition based on coding mode, and in code efficiency, tool has great advantage; Antigen copies generation initial antibodies to it after generating.

Step b antibody cloning.Increase the scale of initial antibodies or the antibody (i.e. outstanding antibody) to antigen-reactive fierceness, to obtaining more outstanding antibody.

Step c Hypermutation.CHIEA constructs a kind of random mixing variation: the genetic fragment (i is a random number, and 1≤i≤m, m represents machine sum) that first Stochastic choice i machine is corresponding; Then carry out random alignment to random two gene position in each genetic fragment, the result of random alignment is one of two schemes below: 1. exchange mutation (Swapmutation), 2. genetic fragment continues to have.

Steps d self-identifying is decoded.After carrying out hyper mutation to initial antibodies, antibody may be transformed into infeasible solutions, decode time will there is deadlock, self-identifying decoding to each processing work identification and be inserted into machine in machining gap, the dynamic conditioning of infeasible solutions to feasible solution can be realized.

Step e data base upgrades.Fitness value (the decoded Makespan of antibody) according to antibody sorts to the antibody of new generation and parent antibody, then preserves outstanding newborn antibody, eliminates poor antibody.

Step f antibody is selected.Antibody select be in data base, choose some outstanding antibody to carry out antibody cloning operation.Consider the counting yield of algorithm and avoid algorithm to be absorbed in local optimum, the antibody of CHIEA is selected to be the selection based on antibody grade of fit, instead of based on the selection of antibody information entropy and antibody concentration in other documents.

The immunoevolution operation that step g circulation step b to step f is all, until obtain the optimal scheduling scheme Ω of optimum antibody and Job Shop problem:

Ω＝{{B _1l,…,B _vl,…,B _e1;C _ll,…,C _v1,…,C _e1;Z _1l,…,Z _v1,…,Z _e1},…,{B _1i,…,B _vi,…,B _ei;C _1i,…,C _vi,…,C _ei;Z _li,…,Z _vi,…,Z _ei},…,{B _1m,…,B _vm,…,B _em;C _lm,…,C _vm,…,C _em;Z _lm,…,Z _vm,…,Z _em}}

Wherein, B _virepresent v ∈ E={1,2 ..., e} workpiece at the i-th ∈ M={1,2 ..., the beginning process time on m} platform machine, C _virepresent the machine time of v workpiece on i-th machine, Z _virepresent the frock setup time of v workpiece on i-th machine;

Step 2: set up machine and integrate as A={A ₁, A ₂..., A _m, evaluation attributes integrate as X={X ₁, X ₂..., X _n; According to the optimal scheduling scheme Ω that step 1 obtains, calculate each evaluation attributes value of each machine, obtain the decision matrix D=(x for bottleneck identification _ij) _{m × n}, x _ijit is a jth evaluation attributes value of the machine of i-th; Evaluation attributes can in machine burden, machinery utilization rate, and active time, enlivens deviation, susceptibility, unit throughput, carries out combination and selects in progress queues length and overall equipment efficiency.comprehensive efficiency of equipment, choose machine burden in the present embodiment, machinery utilization rate, active time is as evaluation attributes.

Step 3: adopt vectorial standardized method by the decision matrix D=(x of step 2 _ij) _{m × n}convert criteria decision matrix R=(r to _ij) _{m × n}:

$r_{\mathrm{ij}}=x_{\mathrm{ij}}/\sqrt{\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}{x}_{\mathrm{ij}}^2}$

Its objective is and various types of evaluation attributes value is converted to nondimensional attribute, attribute can be compared mutually.

Step 4: at decision matrix D=(x _ij) _{m × n}in, for certain attribute X _jproperty value x _ijgap larger, then this attribute role in decision making package is larger, otherwise then less.Information entropy is the tolerance of the information degree of disorder, and information entropy is larger, and the degree of disorder of information is higher, and its value is less, and the unordered degree of system is less, therefore adopts the degree of order and the effectiveness thereof of information entropy evaluation system information.This step adopts entropy assessment determination evaluation attributes weight, solves evaluation attributes weight vectors W=(ω ₁, ω ₂..., ω _j..., ω _n), entropy assessment can eliminate the artificial interference that each index weights calculates as far as possible, makes evaluation result more realistic.Its calculation procedure is as follows:

Step 4.1: by the standardization of each evaluation attributes value, wherein the nondimensionalization property value p of lower i-th machine of a jth attribute _ijfor:

$p_{\mathrm{ij}}=x_{\mathrm{ij}}/\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}{x}_{\mathrm{ij}}$

Step 4.2: the entropy E calculating a jth evaluation attributes _j:

$E_j=-\mathrm{\&delta;}\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}{p}_{\mathrm{ij}}\ln p_{\mathrm{ij}}$

Wherein δ represents a constant, and δ=1/lnm (ln is natural logarithm), ensure that 0≤E _j≤ 1;

Step 4.3: the errored message degree d calculating a jth attribute _j:

d _j＝1-E _j

Work as d _jtime larger, attribute is more important.

Step 4.4: calculating weight is normalized to errored message degree:

When decision maker does not have the preference between evaluation attributes, can think that this n evaluation attributes have identical preference according to uncertain theory, then:

${\mathrm{\&omega;}}_j=d_j/\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{d}_j$

When decision maker has the preference between evaluation attributes:

${\mathrm{\&omega;}}_j={\mathrm{\&lambda;}}_j{d}_j/\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_j{d}_j$

Wherein λ _jfor the subjective weight of decision maker;

In the present embodiment, think that decision maker does not have the preference of attribute.

Step 5: the evaluation attributes weight inclusive criteria decision matrix R=(r that step 4 is obtained _ij) _{m × n}in, obtain weighting standard decision matrix V=(v _ij) _{m × n}=(ω _jr _ij) _{m × n};

Step 6: the positive and negative ideal solution A determining evaluation attributes ⁺and A ^-:

$A^{+}=\{\left(\underset{i}{\max }{v}_{\mathrm{ij}}\right|j\&Element;J),\left(\underset{i}{\min }{v}_{\mathrm{ij}}\right|j\&Element;J^{\&prime;})|i=M\}=\{v_1^{+},v_2^{+},...,v_j^{+},...,v_n^{+}\}$

$A^{-}=\{\left(\underset{i}{\min }{v}_{\mathrm{ij}}\right|j\&Element;J),\left(\underset{i}{\max }{v}_{\mathrm{ij}}\right|j\&Element;J^{\&prime;})|i\&Element;M\}=\{v_1^{-},v_2^{-},...,v_j^{-},...,v_n^{-}\}$

Wherein J is the set of profit evaluation model evaluation attributes, and J ' is the set of cost type evaluation attributes; For profit evaluation model evaluation attributes, property value is the bigger the better, and for cost type evaluation attributes, property value is the smaller the better.

Step 7: tie up Euclid apart from the distance between each machine of calculating and positive ideal solution and minus ideal result by n: every platform machine to the distance of positive ideal solution is:

$S_{i^{+}}=\sqrt{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{(v_{\mathrm{ij}}-v_j^{+})}^2}$

Every platform machine to the distance of minus ideal result is:

$S_{i^{-}}=\sqrt{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{(v_{\mathrm{ij}}-v_j^{-})}^2}$

Step 8: calculate every platform machine and positive ideal solution A ⁺approach degree C _i:

$C_i=S_{i^{-}}/(S_{i^{+}}+S_{i^{-}}),0<C_i<1$

Obtain approach degree C _imaximum machine is the bottleneck of job shop.

Bottleneck identification results contrast is in table 1.

The results contrast of table 1 more than attribute bottleneck identification method (MABI) and SBD method and orthogonal test bottleneck identification method

Note: add * and represent that bottleneck identification method is not identical with shifting bottleneck recognition result herein; Add * * and represent that bottleneck identification method is not identical with orthogonal experiment recognition result herein;

Can find out that the bottleneck identification result of bottleneck identification method in this paper and moving bottleneck method of identification coincide rate up to 95.8% from the numerical results of table 1, also reach 75% with orthogonal experiment recognition result rate of coincideing.This method have employed machine and on average continues active time as one of index in example, and moving bottleneck method of identification is based on average active time index and the bottleneck identification method carried out improving and proposed.Consider that two kinds of methods are all relevant to and average continue active time factor, this method and moving bottleneck method of identification recognition result similarity relatively high.In addition, it might not be system bottleneck that numerical results demonstrates the high machine of load, as example LA21, LA22, LA27.

Three kinds of bottleneck identification methods have identified different bottlenecks, and being not quite similar of recognition result describes the impact of bottleneck definition on bottleneck identification result.Bottleneck is enlivened the bottleneck machine that duration the longest machine definitions is system by moving bottleneck method of identification, and orthogonal experiment considers the impact of machine on system throughput, is the bottleneck machine of system by the factor definition had the greatest impact for test index.The present invention considers the bottleneck Synthetic Decision Method of multiple bottleneck characteristic attribute, is different from the classic method of the definition identification bottleneck according to machine single factors, is a kind of method of comprehensive evaluation machine.

Claims (2)

1. consider a job shop bottleneck identification method for multidimensional feature attribute of machine, it is characterized in that: comprise the following steps:

Step 1: the optimal scheduling scheme Ω adopting optimized algorithm determination job shop:

Ω＝{{B ₁₁,…,B _v1,…,B _e1;C ₁₁,…,C _v1,…,C _e1;Z ₁₁,…,Z _v1,…,Z _e1},…,{B _1i,…,B _vi,…,B _ei;C _li,…,C _vi,…,C _ei;Z _li,…,Z _vi,…,Z _ei},…,{B _1m,…,B _vm,…,B _em;C _lm,…,C _vm,…,C _em;Z _lm,…,Z _vm,…,Z _em}}

Wherein, B _virepresent v ∈ E={1,2 ..., e} workpiece at the i-th ∈ M={1,2 ..., the beginning process time on m} platform machine, C _virepresent the machine time of v workpiece on i-th machine, Z _virepresent the frock setup time of v workpiece on i-th machine;

Step 2: set up machine and integrate as A={A ₁, A ₂..., A _m, evaluation attributes integrate as X={X ₁, X ₂..., X _n; According to the optimal scheduling scheme Ω that step 1 obtains, calculate each evaluation attributes value of each machine, obtain the decision matrix D=(x for bottleneck identification _ij) _{m × n}, x _ijit is a jth evaluation attributes value of the machine of i-th;

Step 3: adopt vectorial standardized method by the decision matrix D=(x of step 2 _ij) _{m × n}convert criteria decision matrix R=(r to _ij) _{m × n}:

$r_{\mathrm{ij}}=x_{\mathrm{ij}}/\sqrt{\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}{x}_{\mathrm{ij}}^2}$

Step 4: adopt entropy assessment determination evaluation attributes weight, evaluation attributes weight vectors W=(ω ₁, ω ₂..., ω _j..., ω _n):

Step 4.1: by the standardization of each evaluation attributes value, wherein the nondimensionalization property value p of lower i-th machine of a jth attribute _ijfor:

$p_{\mathrm{ij}}=x_{\mathrm{ij}}/\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}{x}_{\mathrm{ij}}$

Step 4.2: the entropy E calculating a jth evaluation attributes _j:

$E_j=-\mathrm{\&delta;}\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}{p}_{\mathrm{ij}}\ln p_{\mathrm{ij}}$

Wherein δ=1/lnm, ensure that 0≤E _j≤ 1;

Step 4.3: the errored message degree d calculating a jth attribute _j:

d _j＝1-E _j

Step 4.4: calculating weight is normalized to errored message degree:

When decision maker does not have the preference between evaluation attributes:

${\mathrm{\&omega;}}_j=d_j/\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{d}_j$

When decision maker has the preference between evaluation attributes:

${\mathrm{\&omega;}}_j={\mathrm{\&lambda;}}_j{d}_j/\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_j{d}_j$

Wherein λ _jfor the subjective weight of decision maker;

Step 5: the evaluation attributes weight inclusive criteria decision matrix R=(r that step 4 is obtained _ij) _{m × n}in, obtain weighting standard decision matrix V=(v _ij) _{m × n}=(ω _jr _ij) _{m × n};

Step 6: the positive and negative ideal solution A determining evaluation attributes ⁺and A ^-:

$A^{+}=\{\left(\underset{i}{\max }{v}_{\mathrm{ij}}\right|j\&Element;J),\left(\underset{i}{\min }{v}_{\mathrm{ij}}\right|j\&Element;J^{\&prime;})|i=M\}=\{v_1^{+},v_2^{+},...,v_j^{+},...,v_n^{+}\}$

$A^{-}=\{\left(\underset{i}{\min }{v}_{\mathrm{ij}}\right|j\&Element;J),\left(\underset{i}{\max }{v}_{\mathrm{ij}}\right|j\&Element;J^{\&prime;})|i\&Element;M\}=\{v_1^{-},v_2^{-},...,v_j^{-},...,v_n^{-}\}$

Wherein J is the set of profit evaluation model evaluation attributes, and J ' is the set of cost type evaluation attributes;

Step 7: tie up Euclid apart from the distance between each machine of calculating and positive ideal solution and minus ideal result by n: every platform machine to the distance of positive ideal solution is:

$S_{i^{+}}=\sqrt{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{(v_{\mathrm{ij}}-v_j^{+})}^2}$

Every platform machine to the distance of minus ideal result is:

$S_{i^{-}}=\sqrt{\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{(v_{\mathrm{ij}}-v_j^{-})}^2}$

Step 8: calculate every platform machine and positive ideal solution A ⁺approach degree C _i:

$C_i=S_{i^{-}}/(S_{i^{+}}+S_{i^{-}}),0<C_i<1$

Obtain approach degree C _imaximum machine is the bottleneck of job shop.

2. a kind of job shop bottleneck identification method considering multidimensional feature attribute of machine according to claim 1, is characterized in that: evaluation attributes comprise machine burden, machinery utilization rate, active time, enlivens deviation, susceptibility, unit throughput, in progress queues length and overall equipment efficiency.comprehensive efficiency of equipment.