CN105069561A - Bidding evaluation expert extraction method and device - Google Patents

Abstract

The invention discloses a bidding evaluation expert extraction method and device. The method includes: extracting bidding evaluation experts of a target number m from a bidding evaluation expert database in a random extraction manner as a first bidding evaluation expert set; sorting the remaining bidding evaluation experts in the bidding evaluation expert database in an ascending manner according to the number of times of being extracted in a preset historical period of time; selecting a plurality of bidding evaluation experts from the sequence with a preset selection percentage value a as a second bidding evaluation expert set, wherein 1>a>0; extracting [b*m] bidding evaluation experts from the second bidding evaluation expert set in a random extraction manner, and adding to the first bidding evaluation expert set, where b is a preset parameter and is greater than 0, and [b*m] represents rounding of a real number b*m; and extracting bidding evaluation experts of the target number m from the first bidding evaluation expert set in a random extraction manner. The method provided by the invention increases the extracted probability of bidding evaluation experts whose number of times of being extracted in the historical period of time is small on the premise of ensuring extraction randomness.

Description

A kind of bidding professor abstracting method and device

Technical field

The application relates to balanced extraction technique field, more particularly, relates to a kind of bidding professor abstracting method and device.

Background technology

Bidding professor refers to the prequalification application documents of inviting and submitting bids and submit to tenderer in government procurement activity and the tender documents professional with certain level that examines or evaluate in accordance with the law.At present, country establishes evaluation expert base, is can choose from evaluation expert base at needs bidding professor.

In order to the justice of the Guarantee item assessment of bids and bidding professor are extracted the equilibrium of chance, be generally to extract bidding professor in prior art in pure mode of randomly drawing.But, the harmony randomly drawing model be mainly reflected in extract number of times enough large when, and the number of times extracting bidding professor in reality does not reach sufficient amount, inventor is by being extracted the statistics of number of times to bidding professor each in reality, find that in evaluation expert base, some bidding professor is few by the number of times drawn, the bidding professor even had was not drawn.

Therefore, existing bidding professor extracts mode to be existed part bidding professor to be extracted probability too low, causes the problem that the fairness of extraction process is on the low side.

Summary of the invention

In view of this, this application provides a kind of bidding professor abstracting method and device, extracting the extraction process fairness that exists of mode problem on the low side for solving existing bidding professor.

To achieve these goals, the existing scheme proposed is as follows:

A kind of bidding professor abstracting method, comprising:

To randomly draw mode, from evaluation expert base, extract destination number m bidding professor, as the first bidding professor collection;

For bidding professor remaining in evaluation expert base, carried out ascending sort according in its preset historical time section by the number of times drawn;

Choose ratio value a with preset, several bidding professor before choosing from the sequence after ascending sort, as the second bidding professor collection, wherein 1>a>0;

To randomly draw mode, concentrate from described second bidding professor and extract [b*m] individual bidding professor, and be added to described first bidding professor collection, wherein b is preset parameter and is greater than 0, and [b*m] expression rounds real number b*m;

To randomly draw mode, concentrate from described first bidding professor and extract destination number m bidding professor, as Output rusults.

Preferably, described to randomly draw mode, from evaluation expert base, extract destination number m bidding professor, comprising:

Carry out randomly ordered to all bidding professor in evaluation expert base;

Choose destination number m bidding professor before in sequence.

Preferably, described preset parameter b is less than 1.

A kind of bidding professor draw-out device, comprising:

First extracting unit, for randomly draw mode, extracts destination number m bidding professor, as the first bidding professor collection from evaluation expert base;

Residue expert sequencing unit, for for bidding professor remaining in evaluation expert base, is carried out ascending sort according in its preset historical time section by the number of times drawn;

Second extracting unit, for choosing ratio value a with preset, several bidding professor before choosing from the sequence after ascending sort, as the second bidding professor collection, wherein 1>a>0;

3rd extracting unit, for to randomly draw mode, concentrate from described second bidding professor and extract [b*m] individual bidding professor, and be added to described first bidding professor collection, wherein b is preset parameter and is greater than 0, and [b*m] expression rounds real number b*m;

4th extracting unit, for randomly draw mode, concentrates from described first bidding professor and extracts destination number m bidding professor, as Output rusults.

Preferably, described first extracting unit comprises:

Whole expert's sequencing unit, for carrying out randomly ordered to all bidding professor in evaluation expert base;

Sequence chooses unit, for choosing destination number m bidding professor before in sequence.

Preferably, described preset parameter b is less than 1.

As can be seen from above-mentioned technical scheme, the bidding professor abstracting method that the embodiment of the present application provides, first to randomly draw mode, destination number m bidding professor is extracted from evaluation expert base, as the first bidding professor collection, further for bidding professor remaining in evaluation expert base, ascending sort is carried out by the number of times drawn according in its preset historical time section, and choose ratio value a with preset, several bidding professor before choosing from the sequence after ascending sort, as the second bidding professor collection, then to randomly draw mode, concentrate from described second bidding professor and extract [b*m] individual bidding professor, and be added to described first bidding professor collection, last to randomly draw mode, concentrate from described first bidding professor and extract destination number m bidding professor, as Output rusults.The method of the application, is ensureing under the prerequisite extracting randomness, improves in historical time section and is extracted probability by what draw the few bidding professor of number of times, improve the harmony of extraction on the whole, ensure that the fairness of extraction process.

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present application or technical scheme of the prior art, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is only the embodiment of the application, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to the accompanying drawing provided.

Fig. 1 is a kind of bidding professor abstracting method process flow diagram disclosed in the embodiment of the present application;

Fig. 2 is a kind of bidding professor draw-out device structural representation disclosed in the embodiment of the present application.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present application, be clearly and completely described the technical scheme in the embodiment of the present application, obviously, described embodiment is only some embodiments of the present application, instead of whole embodiments.Based on the embodiment in the application, those of ordinary skill in the art are not making the every other embodiment obtained under creative work prerequisite, all belong to the scope of the application's protection.

In order to solve in reality in bidding professor extraction process, adopt that pure to randomly draw the extraction that model brings unbalanced, part bidding professor is drawn the few problem of number of times, the embodiment of the present application provides a kind of bidding professor abstracting method, see Fig. 1, Fig. 1 a kind of bidding professor abstracting method process flow diagram disclosed in the embodiment of the present application.

As shown in Figure 1, the method comprises:

Step S100, to randomly draw mode, from evaluation expert base, extract destination number m bidding professor, as the first bidding professor collection;

Suppose total n bidding professor in evaluation expert base, and the bidding professor number required for the Bid Evaluation of certain project is m, then adopt the mode randomly drawed, from evaluation expert base, randomly draw m bidding professor, form the first bidding professor collection by this m bidding professor.

Step S110, for bidding professor remaining in evaluation expert base, carried out ascending sort according in its preset historical time section by the number of times drawn;

Evaluation expert base after extracting for step S100, its residue bidding professor number is n-m.For remaining this n-m bidding professor, carried out ascending sort according in its preset historical time section by the number of times drawn.Here, preset historical time section can be set by user, such as the previous year, one month etc.In the sequence obtained after ascending sort, drawn the few bidding professor of number of times in historical time section, be positioned at the front end of sequence.

Step S120, choose ratio value a with preset, several bidding professor before choosing from the sequence after ascending sort, as the second bidding professor collection;

Particularly, the preset ratio value a that chooses is the parameter that user presets, and meets 1>a>0.A can regard equilibrium as and improve coefficient, hereafter can improve degree to the size of a value to equilibrium and analyze.In this step, for the sequence that previous step obtains, from sequence, choose a certain amount of bidding professor to choose ratio value a, and these bidding professor chosen are in sequence from first bidding professor, and several bidding professor of continuous print.

Citing as, have 100 bidding professor in sequence, be numbered No. 1 bidding professor, No. 2 bidding professor respectively according to clooating sequence ... ..100 number bidding professor.The preset ratio value a that chooses is 40%, then choose 1-40 bidding professor in sequence.

Step S130, to randomly draw mode, concentrate from described second bidding professor and extract [b*m] individual bidding professor, and be added to described first bidding professor collection;

Wherein b is preset parameter and is greater than 0, and [b*m] expression rounds real number b*m.For the second bidding professor collection that previous step obtains, according to randomly drawing mode, therefrom extract [b*m] individual bidding professor, and it is concentrated to join the first bidding professor, composition intersection.From defining above, for the first bidding professor collection newly obtained, wherein the number of bidding professor is m+ [b*m].

Step S140, to randomly draw mode, from described first bidding professor concentrate extract destination number m bidding professor, as Output rusults.

Finally, for the first bidding professor collection that previous step obtains, therefrom extract m bidding professor, Output rusults the most final to randomly draw mode.

The bidding professor abstracting method that the embodiment of the present application provides, first to randomly draw mode, destination number m bidding professor is extracted from evaluation expert base, as the first bidding professor collection, further for bidding professor remaining in evaluation expert base, ascending sort is carried out by the number of times drawn according in its preset historical time section, and choose ratio value a with preset, several bidding professor before choosing from the sequence after ascending sort, as the second bidding professor collection, then to randomly draw mode, concentrate from described second bidding professor and extract [b*m] individual bidding professor, and be added to described first bidding professor collection, last to randomly draw mode, concentrate from described first bidding professor and extract destination number m bidding professor, as Output rusults.The method of the application, is ensureing under the prerequisite extracting randomness, improves in historical time section and is extracted probability by what draw the few bidding professor of number of times, improve the harmony of extraction on the whole, ensure that the fairness of extraction process.

Optionally, in order to ensure the randomness extracted, preset parameter b preferably can be chosen as the number being less than 1, and such as 0.5 etc.

From extraction mode above, the bidding professor of the present embodiment extracts mode and ensure that under the prerequisite extracting randomness, improves and extracts harmony.

First, for randomness:

As seen from the above, the present embodiment extraction process have invoked three random functions.It is the bidding professor that follow-up extraction provides maximum quantity that the first step calls random function, and this ensure that the randomness of extraction greatly.Second step is drawn within the scope of the less bidding professor of number of times in historical time section calls random function, also pursuit balanced basis on, to a certain degree ensure that randomness.3rd step, concentrates and calls random function taking into account randomness and balanced the first bidding professor newly, determine final bidding professor, further ensure that extraction randomness.

Secondly, for harmonious:

Balanced improvement is mainly reflected on step S110-S130, is namely extracted within the scope of the less bidding professor of number of times in historical time section and extracts a certain amount of bidding professor, and it is concentrated to integrate with the first bidding professor.Because first extract the order pursued is randomness, next is only harmony, therefore can select for threshold parameter b the numerical value being less than 1, such as 0.5, and then extracts m/2 bidding professor, and it is concentrated to add the first bidding professor.Both improve and extracted the selected probability of the less expert of number of times, greater impact had not been caused to m the bidding professor that first step is selected at random again.

In another embodiment of the application, to above-mentioned steps S100, to randomly draw mode, the process extracting destination number m bidding professor from evaluation expert base is introduced.

Particularly, this extraction process can be subdivided into following two sub-steps:

A, carry out randomly ordered to all bidding professor in evaluation expert base;

B, to choose in sequence before destination number m bidding professor.

Certainly, above-mentioned is only the one of random selecting mode, can also carry out the extraction of bidding professor in other random selecting mode.

Next, first the present embodiment proves in single extraction process, extract the less expert of number of times in improved model (also namely according to the extraction mode of the application) to draw likelihood ratio high in master mould (also namely existing pure randomly draw mode), and provide equilibrium and improve coefficient a to the adjusting range extracting probability.On this basis, prove in one-period, draw certain expert that number of times is always less, his probability of being drawn is improved at improved model, and the cycle of brief analysis improved model balanced improvement degree.

1, single equilibrium improves

1) the balanced improvement of single proves

If x extracts the less bidding professor of number of times in certain historical time section, it is drawn probability in master mould is p ₁, be p by the probability drawn in improved model ₂, existing quantitative test p ₁and p ₂quantitative relation.

Can be drawn by abstracting method:

$p_1=\frac{m}{n}$

$\begin{array}{c}{p}_2=\frac{m}{n}\·\frac{m}{m+\[bm\]}+\frac{n-m}{n}\·\frac{\[bm\]}{a(n-m)}\·\frac{m}{m+\[bm\]}\\ =\frac{m}{n}\·\frac{m}{m+\[bm\]}\·\left(m+\frac{\[bm\]}{a}\right)\end{array}$

It should be noted that, at p ₂computing formula in, when consider the bidding professor x first step selected first bidding professor collection, the bidding professor number that second step extracts scope is:

The probability that bidding professor x is selected in the second bidding professor collection is:

$\frac{\[bm\]}{a(n-m)}\≤1$

Or

In order to ensure that conclusion has ubiquity, above-mentioned we choose less probability

Can find out, to prove p ₂>p ₁, the following condition of a demand fulfillment:

Also be

And owing to having defined 1>a>0 in above-described embodiment, therefore above-mentioned inequality perseverance is set up, thus demonstrate p ₂>p ₁.

2) single equilibrium improves degree analyzing

The same, if x is certain extract the less bidding professor of number of times, he is p by the probability drawn in master mould ₁, be p by the probability drawn in improved model ₂, prove p above ₂>p ₁, this bidding professor of analysis draws the amplitude that probability is enhanced now, i.e. the balanced improvement amplitude I of single ₁.

$I_1=\frac{p_2-p_1}{p_1}$

Because the extraction strategy of improved model is when carrying out second step harmony and extracting, the bidding professor number that extraction scope comprises is:

S has two kinds of possible values, and this extracts the less bidding professor of number of times in improved model by the Probability p drawn by causing ₂there are two kinds of values possibilities, discuss respectively below.

A) as a (n-m)=[bm]:

$\begin{array}{c}{p}_2=\frac{m}{n}\·\frac{m}{m+\[bm\]}+\frac{n-m}{n}\·\frac{\[bm\]}{\[bm\]}\·\frac{m}{m+\[bm\]}\\ =\frac{m}{m+\[bm\]}\end{array}$

$p_1=\frac{m}{n}$

$\begin{array}{c}{I}_1=\frac{p_2-p_1}{p_1}\\ =\frac{\frac{m}{m+\[bm\]}-\frac{m}{n}}{\frac{m}{n}}\\ =\frac{n}{m+\[bm\]}-1\end{array}$

Due to a (n-m)=[bm], therefore to prove I ₁>0, only need prove n>m+a (n-m), also namely prove n>m, and obviously this relation sets up, therefore above-mentioned I ₁perseverance is greater than 0.

B) as a (n-m) > [bm]:

$p_1=\frac{m}{n}$

$\begin{array}{c}{p}_2=\frac{m}{n}\·\frac{m}{m+\[bm\]}+\frac{n-m}{n}\·\frac{\[bm\]}{a(n-m)}\·\frac{m}{m+\[bm\]}\\ =\frac{m}{n}\·\frac{1}{m+\[bm\]}\·\left(m+\frac{\[bm\]}{a}\right)\end{array}$

Therefore, above-mentioned I ₁perseverance is greater than 0.

Following table 1 gives when b value is 0.5, and the minimum single that equilibrium improves corresponding to the different value of coefficient a improves amplitude:

Equilibrium improves factor alpha	Minimum single improves amplitude
		10％	300％
20％	133.33％
		30％	77.78％
40％	50％
		50％	33.34％
60％	22.22％
		70％	129％
80％	8.33％
		90％	3.70％
100％	0

Table 1

As can be seen from the above table, along with equilibrium improves the increase of coefficient, extract the increase rate that the less bidding professor of number of times draws probability and reduce gradually.When equilibrium improve factor alpha get 10% time, extract the less bidding professor of number of times in improved model, drawn that likelihood ratio puns random mold is minimum can improve 3 times.When equilibrium improve factor alpha get 100% time, in fact do not make harmonious adjustment, in this case to be drawn probability identical with puns random mold, do not improve.

User can according to the needs of actual conditions, and the equilibrium that autonomous adjustment α realizes expecting improves degree.

2, cycle equilibrium improves

Considering that expert's equilibrium of assessment of bids history randomly draws model with the number of times that (such as a season, 1 year) expert in one-period is extracted is the harmony that foundation adjusts expert's extraction.Demonstrate above in single extraction process, use improved model can improve the extraction probability of the less expert of extraction number of times of last cycle, we analyze further and draw certain always less expert of number of times below, in one-period, he is drawn the change of probability in improved model.

1) cycle equilibrium improvement proves

Known x extracts the always less expert of number of times in this cycle, he by the probability drawn is in puns random mold in improved model by the probability drawn be demonstrate p above ₂>p ₁.

If this expert randomly draws single in model at puns random mold and equilibrium and is not respectively p ' by the probability drawn ₁, p ' ₂, obviously,

p′ ₁＝1-p ₁，

p′ ₂＝1-p ₂，

And p ' ₁>p ' ₂.

Be located in one-period, it is k time that bidding professor extracts number of times, then this bidding professor x is not all respectively (p by the probability drawn at k time in pure random, balanced two class models ₁') ^k(p ₂') ^k.

Because f (x)=x ^kthe monotone increasing when k>1, x>0, so there is (p ₁') ^k> (p ₂') ^k.

So, this bidding professor x is designated as p by the probability drawn respectively in one-period in pure random, balanced two class models _s, p _j,

p _S＝1-(p ₁') ^k，

p _J＝1-(p ₂') ^k，

Obviously, p _s<p _j.

Also the equilibrium namely demonstrating the consideration assessment of bids history that the application proposes is randomly drawed model and is improve and extract the less bidding professor of number of times in one-period by the probability drawn.

2) cycle equilibrium improves degree analyzing

Improve degree for weighing the equilibrium of improved model in one-period, we introduce an index: cycle equilibrium improves multiple I ₂, represent and extract number of times expert always less in one-period, he is drawn probability and is drawn the multiple of probability in puns random mold after improvement in model.

The same, be located in one-period, it is k time that bidding professor extracts number of times, then this bidding professor x is respectively p by the probability (at least by the probability drawn once) drawn in one-period in pure random, balanced two class models _s=1-(p ₁') ^k, p _j=1-(p ₂') ^k, so, cycle equilibrium improves multiple I ₂computing formula as follows:

$I_2=\frac{p_J}{p_S}=\frac{1-{\left({p_2}^{\&prime;}\right)}^k}{1-{\left({p_1}^{\&prime;}\right)}^k}=\frac{1-{(1-p_2)}^k}{1-{(1-p_1)}^k}$

Substitute into p ₁, p ₂, wherein p ₂value due to improved model second step harmony extract time scope be there are two kinds of possibility values, therefore, I ₂computing formula also divide the following two kinds situation:

A) as a (n-m)=[bm]:

B) as a (n-m) > [bm]:

Cycle equilibrium improves multiple I ₂value depending on concrete extraction situation, namely improved model to extract number of times less expert draw in one-period probability improvement multiple can along with candidate expert number n, extract number m and equilibrium improve the difference of coefficient a, b and change.Be 20 below in conjunction with Candidate Set expert number, extract expert's number be respectively 1,5,10 3 kinds of concrete conditions and parameter b be 0.5, the simple declaration cycle improves multiple.

The equilibrium of table 2 cycle improves multiple (n=20, m=1)

The equilibrium of table 3 cycle improves multiple (n=20, m=5)

The equilibrium of table 4 cycle improves multiple (n=20, m=10)

Wherein, in above-mentioned 3 tables, equilibrium adds * expression a before improving the value of coefficient a does not work, and the scope namely randomly drawing the extraction of second step harmony is

In sum, cycle equilibrium improves multiple and is all more than or equal to 1, illustrates that in balanced extraction model, expert is high in puns random mold by the likelihood ratio drawn in one-period.And when n, m mono-timing, along with the increase extracting number of times k, cycle equilibrium improves multiple and reduces gradually; When n, k mono-timing, along with the increase extracting number m, cycle equilibrium improves multiple and also reduces gradually.

Be described the bidding professor draw-out device that the embodiment of the present application provides below, bidding professor draw-out device described below can mutual corresponding reference with above-described bidding professor abstracting method.

See Fig. 2, Fig. 2 a kind of bidding professor draw-out device structural representation disclosed in the embodiment of the present application.

As shown in Figure 2, this device comprises:

First extracting unit 21, for randomly draw mode, extracts destination number m bidding professor, as the first bidding professor collection from evaluation expert base;

Residue expert sequencing unit 22, for for bidding professor remaining in evaluation expert base, is carried out ascending sort according in its preset historical time section by the number of times drawn;

Second extracting unit 23, for choosing ratio value a with preset, several bidding professor before choosing from the sequence after ascending sort, as the second bidding professor collection, wherein 1>a>0;

3rd extracting unit 24, for to randomly draw mode, concentrate from described second bidding professor and extract [b*m] individual bidding professor, and be added to described first bidding professor collection, wherein b is preset parameter and is greater than 0, and [b*m] expression rounds real number b*m;

4th extracting unit 25, for randomly draw mode, concentrates from described first bidding professor and extracts destination number m bidding professor, as Output rusults.

Optionally, wherein the first extracting unit can comprise:

Whole expert's sequencing unit, for carrying out randomly ordered to all bidding professor in evaluation expert base;

Sequence chooses unit, for choosing destination number m bidding professor before in sequence.

Optionally, in order to ensure the randomness extracted, preset parameter b can select the numerical value being less than 1.

The bidding professor draw-out device that the embodiment of the present application provides, first to randomly draw mode, destination number m bidding professor is extracted from evaluation expert base, as the first bidding professor collection, further for bidding professor remaining in evaluation expert base, ascending sort is carried out by the number of times drawn according in its preset historical time section, and choose ratio value a with preset, several bidding professor before choosing from the sequence after ascending sort, as the second bidding professor collection, then to randomly draw mode, concentrate from described second bidding professor and extract [b*m] individual bidding professor, and be added to described first bidding professor collection, last to randomly draw mode, concentrate from described first bidding professor and extract destination number m bidding professor, as Output rusults.The device of the application, is ensureing under the prerequisite extracting randomness, improves in historical time section and is extracted probability by what draw the few bidding professor of number of times, improve the harmony of extraction on the whole, ensure that the fairness of extraction process.

Finally, also it should be noted that, in this article, the such as relational terms of first and second grades and so on is only used for an entity or operation to separate with another entity or operational zone, and not necessarily requires or imply the relation that there is any this reality between these entities or operation or sequentially.And, term " comprises ", " comprising " or its any other variant are intended to contain comprising of nonexcludability, thus make to comprise the process of a series of key element, method, article or equipment and not only comprise those key elements, but also comprise other key elements clearly do not listed, or also comprise by the intrinsic key element of this process, method, article or equipment.When not more restrictions, the key element limited by statement " comprising ... ", and be not precluded within process, method, article or the equipment comprising described key element and also there is other identical element.

In this instructions, each embodiment adopts the mode of going forward one by one to describe, and what each embodiment stressed is the difference with other embodiments, between each embodiment identical similar portion mutually see.

To the above-mentioned explanation of the disclosed embodiments, professional and technical personnel in the field are realized or uses the application.To be apparent for those skilled in the art to the multiple amendment of these embodiments, General Principle as defined herein when not departing from the spirit or scope of the application, can realize in other embodiments.Therefore, the application can not be restricted to these embodiments shown in this article, but will meet the widest scope consistent with principle disclosed herein and features of novelty.

Claims (6)

1. a bidding professor abstracting method, is characterized in that, comprising:

To randomly draw mode, from evaluation expert base, extract destination number m bidding professor, as the first bidding professor collection;

For bidding professor remaining in evaluation expert base, carried out ascending sort according in its preset historical time section by the number of times drawn;

Choose ratio value a with preset, several bidding professor before choosing from the sequence after ascending sort, as the second bidding professor collection, wherein 1>a>0;

To randomly draw mode, concentrate from described second bidding professor and extract [b*m] individual bidding professor, and be added to described first bidding professor collection, wherein b is preset parameter and is greater than 0, and [b*m] expression rounds real number b*m;

To randomly draw mode, concentrate from described first bidding professor and extract destination number m bidding professor, as Output rusults.

2. method according to claim 1, is characterized in that, described to randomly draw mode, extracts destination number m bidding professor, comprising from evaluation expert base:

Carry out randomly ordered to all bidding professor in evaluation expert base;

Choose destination number m bidding professor before in sequence.

3. method according to claim 1 and 2, is characterized in that, described preset parameter b is less than 1.

4. a bidding professor draw-out device, is characterized in that, comprising:

First extracting unit, for randomly draw mode, extracts destination number m bidding professor, as the first bidding professor collection from evaluation expert base;

Residue expert sequencing unit, for for bidding professor remaining in evaluation expert base, is carried out ascending sort according in its preset historical time section by the number of times drawn;

Second extracting unit, for choosing ratio value a with preset, several bidding professor before choosing from the sequence after ascending sort, as the second bidding professor collection, wherein 1>a>0;

3rd extracting unit, for to randomly draw mode, concentrate from described second bidding professor and extract [b*m] individual bidding professor, and be added to described first bidding professor collection, wherein b is preset parameter and is greater than 0, and [b*m] expression rounds real number b*m;

4th extracting unit, for randomly draw mode, concentrates from described first bidding professor and extracts destination number m bidding professor, as Output rusults.

5. device according to claim 4, is characterized in that, described first extracting unit comprises:

Whole expert's sequencing unit, for carrying out randomly ordered to all bidding professor in evaluation expert base;

Sequence chooses unit, for choosing destination number m bidding professor before in sequence.

6. the device according to claim 4 or 5, is characterized in that, described preset parameter b is less than 1.