CN105894240A - Parameterization-based shift arrangement system and method for flexible service working system - Google Patents

Abstract

The invention discloses a parameterization-based shift arrangement system and method for a flexible service working system. The system comprises a user and staff management module, a historical same-period data management module, an algorithm parameter management module, a shift arrangement time management module, a unit time definition management module, a window number generation module, a newly established shit arrangement module, and a shift arrangement management module. In addition, the invention also provides a parameterization-based shift arrangement method for a flexible service working system. According to the shift arrangement system, a shift arrangement table can be obtained based on prediction and thus a shift arrangement result is generated. The system can carry out window working staff registration restriction, monitoring of various states, and a data analysis, thereby realizing the management function of the window working staff effectively. The window working staff can ask for leave or carry out shift changing directly on the system; and after successful application, the system can adjust a shift arrangement plan and inquire various statistic data of the window working staff, so that the basis can be provided for science shift arrangement.

Description

Based on parameterized elastic services Workforce Management and method

Technical field

The present invention relates to a kind of based on parameterized elastic services Workforce Management and method.

Background technology

In existing window services, the time management of staff uses more traditional mode: open how many windows Mouthful, how open window, completely by manually arranging, sets up staff's shift report.Due to economical swift and violent Exhibition, the project of window services day by day increases, and quantity of service quickly increases, service request is the highest, and existing The open quantity of service Room window and window banging personnel be relatively fixed.If the method not taking science is open Number of windows adjusts staff's configuration, only relies on the service ability that service Room window banging personnel are limited at present And job performance, the service Room is in flow of the people peak period, and the waiting time of client certainly will increase, efficiency of service To give a discount, customer satisfaction with services also will decrease；And service the Room at flow of the people low-valley interval, window Service Source seems again idle and wastes.From the analysis of management and service demand angle, the service Room is urgently to be resolved hurrily Key problem be for customer demand, to carry out rational window and arrange and suitable personnel depaly.

Although occurring in that some scheduling methods at present, such as Patent No. 200810167327.2, entitled one Plant Workforce Management and the Chinese patent of method, but the actual demand of window services industry can not be met.

Summary of the invention

The purpose of the present invention is contemplated to solve the problems referred to above, it is provided that a kind of motility and mobility strong, adapts to Scope wide based on parameterized elastic services scheduling method, it is possible to prediction window number, and being had The result of arranging an order according to class and grade being strict with, improves the enthusiasm of window banging personnel, improves the satisfaction of client.

To achieve these goals, the present invention adopts the following technical scheme that a kind of based on parameterized elastic clothes Business duty scheduling method, comprises the steps: to predict the subsequent work time according to the history business datum same period The window number of Duan Kaifang, the minimum window number that the subsequent work time period opens is drawn by equation below:

$\begin{array}{c}{W}_q=\frac{L_q}{\mathrm{\λ}}=\frac{{\left(c\mathrm{\ρ}\right)}^c\mathrm{\ρ}}{c!{(1-\mathrm{\ρ})}^2}{P}_0=\\ \frac{{\left(c\mathrm{\ρ}\right)}^c\mathrm{\ρ}}{c!{(1-\mathrm{\ρ})}^2}{\[\underset{k=0}{\overset{c-1}{\Σ}}\frac{{\left(c\mathrm{\ρ}\right)}^k}{k!}+\frac{{\left(c\mathrm{\ρ}\right)}^c}{c!(1-\mathrm{\ρ})}\]}^{-1}.\end{array}---\left(1\right)$

Wherein, W_qRepresent client's average latency；C is open minimum window number needed for representing the service Room；Represent service intensity；λ represents the average client's number arriving the service Room per hour；μ represents average every Client's number that one window can service per hour；

t₀Representing the average latency upper limit that client can tolerate, for formula (1), calculation procedure is as follows:

Step 1: order

Step 2: calculate W according to formula (1)_qIf, W_q≤t₀, then terminating algorithm, c now is required Minimum window number；If W_q＞ t₀Then go to step 3；

Step 3: make c=c+1, goes to step 2.

According to window banging personnel's scheduling data of revised prediction window number data genaration each time period, Window banging personnel's scheduling data is drawn by following object function；

Window banging personnel are at the time parameter of window services: the staff of window is once work the most extremely Few stream time section is startmintimeslot, is abbreviated as ST, and default value is 1；Take again after rest The minimum service time continuously of business, number was middlemintimeslot, was abbreviated as MI, and default value is 3, maximum Service time, number was maxtimeslot continuously, was abbreviated as MA, and default value is 5；Window banging personnel's morning or Section is endmintimeslot the minimum working time in the afternoon, is abbreviated as EN, and default value is 2；

Hop count parametrization time shared by whole day；Time shared by the morning, hop count is morningtimeslot, is abbreviated as MO, Default value is 14；Time shared by noon, hop count is noontimeslot, is abbreviated as NO, and default value is 8；Afternoon Time shared, hop count is afternoontimeslot, is abbreviated as AF, and default value is 12；

Between window banging personnel, within a cycle, net cycle time length maximum disparity parameter turns to Allowinterval, is abbreviated as AL, and default value is 4；

Object function is:

$\left\{\begin{array}{c}{\mathrm{minz}}_1=\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{ijk}+(NO-1)\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{T}{\Σ}}{x}_{ij(MO+1)};\\ {\mathrm{minz}}_2=n.\end{array}\right.$

Wherein: T represents plan phase total natural law of arranging an order according to class and grade；N represents the sum of window banging personnel；c_jkExpression is arranged an order according to class and grade The minimum window number needed for the customer demand for service jth sky k moment is met in the plan phase；Decision variable x_ijkRepresent Certain window banging personnel every day in the duty of window banging period, when the window banging personnel day some time When being engraved in window banging, value is 1, is otherwise 0；

Constraints:

$(ST-1)(1-x_{ij,k})(1-x_{ij,k+1})x_{ij,k+2}\≤\underset{n=k+3}{\overset{k+1+ST}{\Σ}}{x}_{ijn},$

$\∀i,j,k=1,2,...,MO-1-ST,MO+2,...,MO+AF-ST;$

$(ST-1)(1-x_{ij1})x_{ij2}\≤\underset{n=3}{\overset{1+ST}{\Σ}}{x}_{ijn},\∀i,j;$

$(ST-1)x_{ij1}\≤\underset{n=2}{\overset{ST}{\Σ}}{x}_{ijn},\∀i,j;$

Above formula ensure that the staff of window works ST time period the most continuously once work, ST≤MI；

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\≤\underset{n=k+3}{\overset{MO}{\Π}}(1-x_{ijn}),$

$\∀i,j,k=1,2,...,MO-3;$

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\≤\underset{n=k+3}{\overset{MO+1+AF}{\Π}}(1-x_{ijn}),$

$\∀i,j,k=MO+2,...,MO+AF-2;$

If above formula is had a rest after 2 time periods after having demonstrate,proved staff's work of window again, then arrive that morning or the same day The most all have a rest in the afternoon；

$(EN-1)x_{ijk}(1-x_{ij,k+1})x_{ij,k+2}\≤\underset{n=k+3}{\overset{k+1+EN}{\Σ}}{x}_{ijn},$

$\∀i,j,k=ST,...,MO-1-EN,MO+1+ST,...,MO+AF-EN;$

Above formula ensure that the staff of window to have a rest after 1 time period and works the most continuously once start working again EN time period, EN≤MI；

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\Π}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO-EN}{\Σ}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\≤(MO-2EN-k-1)\underset{l=k+2+EN}{\overset{k+1+MI}{\Π}}{x}_{ijl},$

$\∀i,j,k=ST,...,MO-2-2EN;$

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\Π}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO+AF+1-EN}{\Σ}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\≤(MO+AF-2EN-k)\underset{l=k+2+EN}{\overset{k+1+MI}{\Π}}{x}_{ijl},$

$\∀i,j,k=MO+1+ST,...,MO+AF-2EN-1;$

Above formula demonstrate,proved the staff of window have a rest after 1 time period start working the most again and be not that morning or Afternoon, last working time section before leaving offices worked MI time period the most continuously；

$\underset{d=k}{\overset{k+MA}{\Σ}}{x}_{ijd}\≤MA,\∀i,j,k=1,2,...,MO-MA,MO+2,...,MO+AF+1-MA$

Above formula ensure that window banging personnel work MA time period the most continuously once work；

$1\≤\[\underset{k=1}{\overset{MO}{\Π}}(1-x_{ijk})+(1-x_{ij,MO+1})+\underset{k=MO+2}{\overset{MO+1+AF}{\Π}}(1-x_{ijk})\]\≤2,\∀i,j$

Above formula ensure that any one sky in window banging person works's day, the i.e. morning, noon, afternoon, only work 2 Individual time period or 1 time period, the darg time is less than 8 hours；

$\frac{1}{2}(x_{ij,MO-1}+x_{ij,MO})\≤1-x_{ij,MO+1},\∀i,j$

If above formula ensure that window banging personnel, certain day noon is on duty, then he (she) within front 2 time periods on duty not In window services, so ensure that noon has the time to have lunch and has time on duty at noon before on duty；

$x_{ijMO}\≤x_{ij,MO-1},x_{ij,MO+1+AF}\≤x_{ij,MO+AF}\∀i,j$

Above formula ensure that window banging personnel's morning or afternoon can not only work 1 time period before leaving offices；

(1-x_{Ij, MO-2})x_{Ij, MO-1}≤x_{Ij, MO},

$(1-x_{ij,MO+AF-1})x_{ij,MO+AF}\≤x_{ij,MO+AF+1}\∀i,j$

Above formula ensure that window banging personnel can only work 2 time periods before leaving offices in the morning or afternoon；

$x_{ijMO}\left(\underset{k=MO+1}{\overset{MO+AF}{\Σ}}{x}_{ijk}\right)=0,\∀j,i,$

If above formula ensure window banging personnel on duty for noon, then afternoon he (she) not at window banging；

$\underset{i=1}{\overset{n}{\Σ}}{x}_{ijk}\≥c_{jk},\∀j,k$

Above formula ensure that jth sky random time section k is operated at full capacity；

$\underset{i=1}{\overset{n}{\Σ}}{x}_{ij,MO+1}=1,\∀j$

Above formula ensure that jth sky have noon and only 1 window banging personnel on duty；

$|\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{ljk}+(NO-1)\underset{j=1}{\overset{T}{\Σ}}{x}_{lj,MO+1}-\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{mjk}-(NO-1)\underset{j=1}{\overset{T}{\Σ}}{x}_{mj,MO+1}|\≤AL,\∀l,m$

Above formula ensures that between window banging personnel, net cycle time length is less than AL time period within a cycle.

Preferably: the history business datum same period includes business item, time of drawing a bill, waiting time, the time of completing Between, accept window and window banging personnel.

Preferably: client average latency W_qLess than 15 minutes, i.e. W_q≤ 0.25 hour.

On the other hand, a kind of based on parameterized elastic services Workforce Management, including:

History contemporaneous data management module, for the record of historical data and management:

Window number generation module, according to the window that the history business datum same period prediction subsequent work time period is open Number, the minimum window number that the subsequent work time period opens is drawn by equation below:

$\begin{array}{c}{W}_q=\frac{L_q}{\mathrm{\λ}}=\frac{{\left(c\mathrm{\ρ}\right)}^c\mathrm{\ρ}}{c!{(1-\mathrm{\ρ})}^2}{P}_0=\\ \frac{{\left(c\mathrm{\ρ}\right)}^c\mathrm{\ρ}}{c!{(1-\mathrm{\ρ})}^2}{\[\underset{k=0}{\overset{c-1}{\Σ}}\frac{{\left(c\mathrm{\ρ}\right)}^k}{k!}+\frac{{\left(c\mathrm{\ρ}\right)}^c}{c!(1-\mathrm{\ρ})}\]}^{-1}.\end{array}---\left(1\right)$

Wherein, W_qRepresent client's average latency；C is open minimum window number needed for representing the service Room；Represent service intensity；λ represents the average client's number arriving the service Room per hour；μ represents average every Client's number that one window can service per hour；

t₀Representing the average latency upper limit that client can tolerate, for formula (1), calculation procedure is as follows:

Step 1: order

Step 2: calculate W according to formula (1)_qIf, W_q≤t₀, then terminating algorithm, c now is required Minimum window number；If W_q＞ t₀Then go to step 3；

Step 3: make c=c+1, goes to step 2.

Newly-built module of arranging an order according to class and grade, according to the window banging of revised prediction window number data genaration each time period Personnel's scheduling data, window banging personnel's scheduling data is drawn by following object function；

Window banging personnel are at the time parameter of window services: the staff of window is once work the most extremely Few stream time section is startmintimeslot, is abbreviated as ST, and default value is 1；Take again after rest The minimum service time continuously of business, number was middlemintimeslot, was abbreviated as MI, and default value is 3, maximum Service time, number was maxtimeslot continuously, was abbreviated as MA, and default value is 5；Window banging personnel's morning or Section is endmintimeslot the minimum working time in the afternoon, is abbreviated as EN, and default value is 2；

Hop count parametrization time shared by whole day；Time shared by the morning, hop count is morningtimeslot, is abbreviated as MO, Default value is 14；Time shared by noon, hop count is noontimeslot, is abbreviated as NO, and default value is 8；Afternoon Time shared, hop count is afternoontimeslot, is abbreviated as AF, and default value is 12；

Between window banging personnel, within a cycle, net cycle time length maximum disparity parameter turns to Allowinterval, is abbreviated as AL, and default value is 4；

Object function is:

$\left\{\begin{array}{c}{\mathrm{minz}}_1=\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{ijk}+(NO-1)\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{T}{\Σ}}{x}_{ij(MO+1)};\\ {\mathrm{minz}}_2=n.\end{array}\right.$

Wherein: T represents plan phase total natural law of arranging an order according to class and grade；N represents the sum of window banging personnel；c_jkExpression is arranged an order according to class and grade The minimum window number needed for the customer demand for service jth sky k moment is met in the plan phase；Decision variable x_ijkRepresent Certain window banging personnel every day in the duty of window banging period, when the window banging personnel day some time When being engraved in window banging, value is 1, is otherwise 0；

Constraints:

$(ST-1)(1-x_{ij,k})(1-x_{ij,k+1})x_{ij,k+2}\≤\underset{n=k+3}{\overset{k+1+ST}{\Σ}}{x}_{ijn},$

$\∀i,j,k=1,2,...,MO-1-ST,MO+2,...,MO+AF-ST;$

$(ST-1)(1-x_{ij1})x_{ij2}\≤\underset{n=3}{\overset{1+ST}{\Σ}}{x}_{ijn},\∀i,j;$

$(ST-1)x_{ij1}\≤\underset{n=2}{\overset{ST}{\Σ}}{x}_{ijn},\∀i,j;$

Above formula ensure that the staff of window works ST time period the most continuously once work, ST≤MI；

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\≤\underset{n=k+3}{\overset{MO}{\Π}}(1-x_{ijn}),$

$\∀i,j,k=1,2,...,MO-3;$

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\≤\underset{n=k+3}{\overset{MO+1+AF}{\Π}}(1-x_{ijn}),$

$\∀i,j,k=MO+2,...,MO+AF-2;$

If above formula is had a rest after 2 time periods after ensure that staff's work of window again, then arrive that morning or ought Afternoon is the most all had a rest；

$(EN-1)x_{ijk}(1-x_{ij,k+1})x_{ij,k+2}\≤\underset{n=k+3}{\overset{k+1+EN}{\Σ}}{x}_{ijn},$

$\∀i,j,k=ST,\mathrm{...},MO-1-EN,MO+1+ST,\mathrm{...},MO+AF-EN;$

Above formula ensure that the staff of window to have a rest after 1 time period and works the most continuously once start working again EN time period, EN≤MI；

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\Π}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO-EN}{\Σ}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\≤(MO-2EN-k-1)\underset{l=k+2+EN}{\overset{k+1+MI}{\Π}}{x}_{ijl},$

$\∀i,j,k=ST,...,MO-2-2EN;$

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\Π}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO+AF+1-EN}{\Σ}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\≤(MO+AF-2EN-k)\underset{l=k+2+EN}{\overset{k+1+MI}{\Π}}{x}_{ijl},$

$\∀i,j,k=MO+1+ST,...,MO+AF-2EN-1;$

Above formula ensure that the staff of window to have a rest after 1 time period to start working the most again and is not that morning Or the last working time section that afternoon is before leaving offices works MI time period the most continuously；

$\underset{d=k}{\overset{k+MA}{\Σ}}{x}_{ijd}\≤MA,\∀i,j,k=1,2,...,MO-MA,MO+2,...,MO+AF+1-MA$

Above formula ensure that window banging personnel work MA time period the most continuously once work；

$1\≤\[\underset{k=1}{\overset{MO}{\Π}}(1-x_{ijk})+(1-x_{ij,MO+1})+\underset{k=MO+2}{\overset{MO+1+AF}{\Π}}(1-x_{ijk})\]\≤2,\∀i,j$

Above formula ensure that any one sky in window banging person works's day, the i.e. morning, noon, afternoon, only work 2 Individual time period or 1 time period, the darg time is less than 8 hours；

$\frac{1}{2}(x_{ij,MO-1}+x_{ij,MO})\≤1-x_{ij,MO+1},\∀i,j$

If above formula ensure that window banging personnel, certain day noon is on duty, then he (she) within front 2 time periods on duty not In window services, so ensure that noon has the time to have lunch and has time on duty at noon before on duty；

$x_{ijMO}\≤x_{ij,MO-1},x_{ij,MO+1+AF}\≤x_{ij,MO+AF}\∀i,j$

Above formula ensure that window banging personnel's morning or afternoon can not only work 1 time period before leaving offices；

(1-x_{Ij, MO-2})x_{Ij, MO-1}≤x_{Ij, MO},

$(1-x_{ij,MO+AF-1})x_{ij,MO+AF}\≤x_{ij,MO+AF+1}\∀i,j$

Above formula ensure that window banging personnel can only work 2 time periods before leaving offices in the morning or afternoon；

$x_{ijMO}\left(\underset{k=MO+1}{\overset{MO+AF}{\Σ}}{x}_{ijk}\right)=0,\∀j,i,$

If above formula ensure window banging personnel on duty for noon, then afternoon he (she) not at window banging；

$\underset{i=1}{\overset{n}{\Σ}}{x}_{ijk}\≥c_{jk},\∀j,k$

Above formula ensure that jth sky random time section k is operated at full capacity；

$\underset{i=1}{\overset{n}{\Σ}}{x}_{ij,MO+1}=1,\∀j$

Above formula ensure that jth sky have noon and only 1 window banging personnel on duty；

$|\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{ljk}+(NO-1)\underset{j=1}{\overset{T}{\Σ}}{x}_{lj,MO+1}-\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{mjk}-(NO-1)\underset{j=1}{\overset{T}{\Σ}}{x}_{mj,MO+1}|\≤AL,\∀l,m$

Above formula ensures that between window banging personnel, net cycle time length is less than AL time period within a cycle.

Preferably: in history contemporaneous data management module, the history business datum same period includes business item, goes out Ticket time, waiting time, deadline, accept window and window banging personnel.

Preferably: in window number generation module, client average latency W_qLess than 15 minutes, i.e. W_q≤ 0.25 hour.

Preferably: include user and staff's benefits module, for record window staff's personal information, window Mouth staff's personal information includes name, job number and telephone number.

Preferably: include that algorithm parameter manages module, for prediction window number and arrange an order according to class and grade needed for algorithm each Plant parameter and carry out typing and dynamically management, including traffic handling time, business weight, queue length, client Between the latency tolerant upper limit, employee's High-effective Service duration, period maximal workload and minimum workload Maximum disparity requires, participate in arranging an order according to class and grade hop count and maximum motor-driven staff number every day time shared by quantity of cadres, noon.

Preferably: segment management module when including arranging an order according to class and grade, when prediction window number, every day is classified the period Concrete start and end time carries out typing and management.

Preferably: include management module of arranging an order according to class and grade, for realization arrange an order according to class and grade person works measure query statistic, data backup, Reduction, system user and rights management.

Elastic services system refers to complete task or the premise of fixing working time length of regulation Under, staff can neatly, independently select the contents such as concrete time or the place of work, to replace The system in unified, fixing commuter time.The implementation of this working system, for establishing good window clothes Business external, strengthen for the humanistic care of window banging personnel, promoting window services efficiency, meaning Far-reaching.

1. improve window banging efficiency, reduce client's transacting business time.The total management system in the service Room is not Only record each item data of the transacting business of client, the most also window banging personnel are carried out performance appraisal." work Make efficiency " be the leading indicator weighing job performance, and improve work efficiency it is crucial that rationally shorten every Business handle duration.And the key factor affecting efficiency is staff's Physiological Psychology stability state, service Ability and service level.Implement elastic services system, core work time and non-core work can be distinguished Time, it is ensured that window banging personnel obtain enough rests and vocational study in the non-core working time, at core Heart working time i.e. business the busiest time period has the energy of abundance to keep optimum state, devotes oneself to work, from And improve job performance.Implement elastic services system, can according to client's flow and the measuring and calculating of portfolio, Open suitable service window, staff provides service with optimal working condition simultaneously, and client can obtain To good service, it is possible to reduce waiting time and transacting business time.

2. alleviate window banging personnel's pressure, meet Batch Processing needs.Window banging makes because of its particularity Pressure source is from many-side.One is that post is fixed, the most constant including job site, working time.Two are Content is single, the business operation relative mechanical of process, repetition.Three is that service standard is unified, discipline execution is tight Lattice, including window mental outlook, operating process, service quality etc..All day is in the face of client, window banging people Member be in for a long time spirit high pressure conditions, once cause anxiety, be sick of etc. negative emotions, easily affect work Efficiency, reduction service level, serious will damage customer relationship.Therefore, set up comprehensive mental pressure to dredge Lead decompression mechanism without benefits.Setting up elastic services system, window banging personnel will have reasonable time Acceptable learning is dredged, reduces pressure and is had a rest, and conscientiously alleviates long-term physical weariness, spiritual high pressure conditions.Carry out Elastic services system, also allow for window banging personnel can planned, have the time to carry out customer profile on backstage Arrangement, amended record client-related information, associated materials are transferred to administration section, effectively meet background work needs.

3. promote post learning and communication, promote professional qualities.The window management degree of becoming more meticulous improves constantly, and The stationarity in window post and the unicity of job content largely affect the lifting of ' professional. Under elastic services system, window personnel can utilize the elastic services time to carry out window service study, Business communication can be carried out with other posies personnel, carry out mutually study, allow more window management " specialized personnel " Change to " versatile person ".

Compared with prior art, there is advantages that Workforce Management energy proposed by the invention Enough obtain shift report by prediction thus generate result of arranging an order according to class and grade.Can be by window banging personnel be logged in restriction Monitoring with various states and data analysis, be effectively realized the management function to window banging personnel, window Staff can ask for leave the most on this system, adjust class, after apply for successfully system adjustable entire row shift plan with Window banging demographic data, arranges an order according to class and grade offer foundation for science.

Accompanying drawing explanation

Fig. 1 is based on parameterized elastic services Workforce Management corn module signal in embodiment Figure.

Fig. 2 is based on parameterized elastic services Workforce Management basic data and secondary mould in embodiment Block schematic diagram.

Fig. 3 is based on parameterized elastic services Workforce Management flow chart in embodiment.

Detailed description of the invention

The present invention arranges an order according to class and grade on the basis of factor at systematic analysis restriction window, application queueing theory and non-linear integer It is a kind of based on parameterized elastic services Workforce Management and method that planning establishes service hall.This row Class's system and method motility and mobility strong, wide accommodation.

Embodiment one: see Fig. 1 to Fig. 3, a kind of based on parameterized elastic services scheduling method, Comprise the steps: the window number open according to the history business datum same period prediction subsequent work time period, under The minimum window number that one working time section opens is drawn by equation below:

$\begin{array}{c}{W}_q=\frac{L_q}{\mathrm{\λ}}=\frac{{\left(c\mathrm{\ρ}\right)}^c\mathrm{\ρ}}{c!{(1-\mathrm{\ρ})}^2}{P}_0=\\ \frac{{\left(c\mathrm{\ρ}\right)}^c\mathrm{\ρ}}{c!{(1-\mathrm{\ρ})}^2}{\[\underset{k=0}{\overset{c-1}{\Σ}}\frac{{\left(c\mathrm{\ρ}\right)}^k}{k!}+\frac{{\left(c\mathrm{\ρ}\right)}^c}{c!(1-\mathrm{\ρ})}\]}^{-1}.\end{array}---\left(1\right)$

Wherein, W_qRepresent client's average latency；C is open minimum window number needed for representing the service Room；Represent service intensity；λ represents the average client's number arriving the service Room per hour；μ represents average every Client's number that one window can service per hour；

t₀Representing the average latency upper limit that client can tolerate, for formula (1), calculation procedure is as follows:

Step 1: order

Step 2: calculate W according to formula (1)_qIf, W_q≤t₀, then terminating algorithm, c now is required Minimum window number；If W_q＞ t₀Then go to step 3；

Step 3: make c=c+1, goes to step 2；

According to window banging personnel's scheduling data of revised prediction window number data genaration each time period, Window banging personnel's scheduling data is drawn by following object function；

Window banging personnel are at the time parameter of window services: the staff of window is once work the most extremely Few stream time section is startmintimeslot (being abbreviated as ST, default value is 1)；Take again after rest The minimum service time continuously of business, number was middlemintimeslot (being abbreviated as MI, default value is 3), Big service time continuously, number was maxtimeslot (being abbreviated as MA, default value is 5)；On window banging personnel Noon or section are endmintimeslot (being abbreviated as EN, default value is 2) the minimum working time in the afternoon.

Hop count parametrization time shared by whole day；Time shared by the morning hop count be morningtimeslot (be abbreviated as MO, Default value is 14)；Time shared by noon, hop count is noontimeslot (being abbreviated as NO, default value is 8)；Under Time shared by noon, hop count is afternoontimeslot (being abbreviated as AF, default value is 12).

Between window banging personnel, within a cycle, net cycle time length maximum disparity parameter turns to Allowinterval (is abbreviated as AL, default value is 4).Object function is as follows:

$\left\{\begin{array}{c}{\mathrm{minz}}_1=\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{ijk}+(NO-1)\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{T}{\Σ}}{x}_{ij(MO+1)};\\ {\mathrm{minz}}_2=n.\end{array}\right.$

Wherein: T represents plan phase total natural law of arranging an order according to class and grade；N represents the sum of window banging personnel；c_jkExpression is arranged an order according to class and grade The minimum window number needed for the customer demand for service jth sky k moment is met in the plan phase；Decision variable x_ijkRepresent Certain window banging personnel every day in the duty of window banging period, when the window banging personnel day some time When being engraved in window banging, value is 1, is otherwise 0；

Constraints is as follows:

Description according to window shifts arrangement and analysis, and combine the definition of decision variable, the constraints of model Mainly have:

$(ST-1)(1-x_{ij,k})(1-x_{ij,k+1})x_{ij,k+2}\≤\underset{n=k+3}{\overset{k+1+ST}{\Σ}}{x}_{ijn},$

$\∀i,j,k=1,2,...,MO-1-ST,MO+2,...,MO+AF-ST;$

$(ST-1)(1-x_{ij1})x_{ij2}\≤\underset{n=3}{\overset{1+ST}{\Σ}}{x}_{ijn},\∀i,j;$

$(ST-1)x_{ij1}\≤\underset{n=2}{\overset{ST}{\Σ}}{x}_{ijn},\∀i,j.$

Above formula ensure that the staff of window works ST (≤MI) the individual time period the most continuously once work；

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\≤\underset{n=k+3}{\overset{MO}{\Π}}(1-x_{ijn}),$

$\∀i,j,k=1,2,...,MO-3;$

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\≤\underset{n=k+3}{\overset{MO+1+AF}{\Π}}(1-x_{ijn}),$

$\∀i,j,k=MO+2,...,MO+AF-2;$

If above formula is had a rest after 2 time periods after ensure that staff's work of window again, then arrive that morning or ought Afternoon is the most all had a rest；

$(EN-1)x_{ijk}(1-x_{ij,k+1})x_{ij,k+2}\≤\underset{n=k+3}{\overset{k+1+EN}{\Σ}}{x}_{ijn},$

$\∀i,j,k=ST,...,MO-1-EN,MO+1+ST,...,MO+AF-EN;$

Above formula ensure that the staff of window to have a rest after 1 time period and works the most continuously once start working again EN (≤MI) the individual time period；

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\Π}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO-EN}{\Σ}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\≤(MO-2EN-k-1)\underset{l=k+2+EN}{\overset{k+1+MI}{\Π}}{x}_{ijl},$

$\∀i,j,k=ST,...,MO-2-2EN;$

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\Π}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO+AF+1-EN}{\Σ}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\≤(MO+AF-2EN-k)\underset{l=k+2+EN}{\overset{k+1+MI}{\Π}}{x}_{ijl},$

$\∀i,j,k=MO+1+ST,...,MO+AF-2EN-1;$

Above formula ensure that the staff of window to have a rest after 1 time period to start working the most again and is not that morning Or the last working time section that afternoon is before leaving offices works MI time period the most continuously；

$\underset{d=k}{\overset{k+MA}{\Σ}}{x}_{ijd}\≤MA,\∀i,j,k=1,2,...,MO-MA,MO+2,...,MO+AF+1-MA$

Above formula ensure that window banging personnel work MA time period the most continuously once work；

$1\≤\[\underset{k=1}{\overset{MO}{\Π}}(1-x_{ijk})+(1-x_{ij,MO+1})+\underset{k=MO+2}{\overset{MO+1+AF}{\Π}}(1-x_{ijk})\]\≤2,\∀i,j$

Above formula ensure that the only work 2 of any one sky (morning, noon, afternoon) in window banging person works's day Time period or 1 time period (the darg time is less than 8 hours)；

$\frac{1}{2}(x_{ij,MO-1}+x_{ij,MO})\≤1-x_{ij,MO+1},\∀i,j$

If above formula ensure that window banging personnel, certain day noon is on duty, then he (she) within front 2 time periods on duty not In window services, so ensure that noon has the time to have lunch and has time on duty at noon before on duty；

$x_{ijMO}\≤x_{ij,MO-1},x_{ij,MO+1+AF}\≤x_{ij,MO+AF}\∀i,j$

Above formula ensure that window banging personnel's morning or afternoon can not only work 1 time period before leaving offices；

(1-x_{Ij, MO-2})x_{Ij, MO-1}≤x_{Ij, MO},

$(1-x_{ij,MO+AF-1})x_{ij,MO+AF}\≤x_{ij,MO+AF+1}\∀i,j$

Above formula ensure that window banging personnel can only work 2 time periods before leaving offices in the morning or afternoon；

$x_{ijMO}\left(\underset{k=MO+1}{\overset{MO+AF}{\Σ}}{x}_{ijk}\right)=0,\∀j,i,$

If above formula ensure window banging personnel on duty for noon, then afternoon he (she) not at window banging；

$\underset{i=1}{\overset{n}{\Σ}}{x}_{ijk}\≥c_{jk},\∀j,k$

Above formula ensure that jth sky random time section k (by historical data, utilizes the theory of statistics and queueing theory to protect Barrier client's waiting time do not surpass 15 minutes estimate window number) operate at full capacity；

$\underset{i=1}{\overset{n}{\Σ}}{x}_{ij,MO+1}=1,\∀j$

Above formula ensure that jth sky have noon and only 1 window banging personnel on duty；

$|\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{ljk}+(NO-1)\underset{j=1}{\overset{T}{\Σ}}{x}_{lj,MO+1}-\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{mjk}-(NO-1)\underset{j=1}{\overset{T}{\Σ}}{x}_{mj,MO+1}|\≤AL,\∀l,m$

Above formula ensure that between window banging personnel, net cycle time length is less than AL time period within a cycle (embodying fair principle).

The history business datum same period includes business item, time of drawing a bill, waiting time, deadline, accepts Window and working window cadre.

Client average latency W_qLess than 15 minutes, i.e. W_q≤ 0.25 hour.

Embodiment two: a kind of based on parameterized elastic services Workforce Management, mainly includes that history is same Phase data management module, window number generation module and newly-built module etc. of arranging an order according to class and grade.

History contemporaneous data management module is for the record of historical data and management:

The window that window number generation module opened according to the history business datum same period prediction subsequent work time period Number.

Extract service Room window every historical data (client arrive the service Room time, etc. the time to be serviced Service time in the Room and the average time of window banging personnel's transacting business is left with finishing thing), during application Between sequence analysis, it was predicted that following a period of time unit time inner segment client arrives the flow of the people in the service Room, profit Go out window by the Theoretical Calculation of queueing theory and offer quantity.Meeting client's waiting time less than 15 minutes (ginsengs Number) on the premise of, client is set to variable waiting time.Working time section can according to local working requirement, Make seasonal adjustment.Client needs to be set as parameter waiting time, can carry out more according to the practical situation of user Change.Open window number is set as parameter, and user can the practical situation of basis be adjusted.Client is arrived A large amount of historical datas of the flow of the people reaching service Room day part are divided into two classes, a class to be dull seasons, and another kind of is prosperous Season.Further according to the attribute of window services industry, the data that client in month arrives service Room flow of the people are divided Being two classes, be divided into busy the such as beginning of the month, the end of month, section remaining time (intermediate period) of month is idle. Calculate each period client in dull season (busy and idle time) again and arrive the average total number of persons in the service Room and the busy season is (busy Time and idle) each period client arrive service the Room average total number of persons, such as day part refers to 8:30-9:30,9:30-10:30 etc..Each period client of recycling dull season (busy and idle time) arrives clothes ARMA model (ARMA) in the average total number of persons binding time sequence analysis in the business Room dopes light Each period client in season (busy and idle time) arrives the number in the service Room；Use the busy season (busy and idle time) Each period client arrives the average total number of persons in the service Room and combines ARMA model in season (SARMA) Each period client doping the busy season (busy and idle time) arrives the number in the service Room.The service of window meets Separate quantum condition entropy, the average customer people serviced in the parameter μ representation unit time of quantum condition entropy Number, we arrive a large amount of historical datas in the service Room and test client, find that client arrives the service Room Rule substantially obeys Poisson distribution, is (can be set to ginseng in 15 minutes in the average latency that client can tolerate Number) hypothesis under, therefore we can utilize M/M/c model (formula (1) the most hereinafter) meter that queueing theory is classical Calculate the window number that each period rationally opens.

The minimum window number that the subsequent work time period opens is drawn by equation below:

$\begin{array}{c}{W}_q=\frac{L_q}{\mathrm{\λ}}=\frac{{\left(c\mathrm{\ρ}\right)}^c\mathrm{\ρ}}{c!{(1-\mathrm{\ρ})}^2}{P}_0=\\ \frac{{\left(c\mathrm{\ρ}\right)}^c\mathrm{\ρ}}{c!{(1-\mathrm{\ρ})}^2}{\[\underset{k=0}{\overset{c-1}{\Σ}}\frac{{\left(c\mathrm{\ρ}\right)}^k}{k!}+\frac{{\left(c\mathrm{\ρ}\right)}^c}{c!(1-\mathrm{\ρ})}\]}^{-1}.\end{array}---\left(1\right)$

Wherein, W_qRepresent client's average latency；C is open minimum window number needed for representing the service Room；Represent service intensity；λ represents the average client's number arriving the service Room per hour；μ represents average every Client's number that one window can service per hour；

t₀Representing the average latency upper limit that client can tolerate, for formula (1), calculation procedure is as follows (owing to c is positive integer and is difficult to be obtained by analysis mode, therefore calculating through the following steps):

Step 1: order(stable state can be reached for queuing process to consider)

Step 2: calculate W according to formula (1)_qIf, W_q≤t₀, then terminating algorithm, c now is required Minimum window number；If W_q＞ t₀Then go to step 3；

Step 3: make c=c+1, goes to step 2；

Business averagely handles time acquisition: when importing new historical data in system, system is carried out newly automatically Business averagely handle the calculating of time.

Prediction window generates, and when user selects to arrange an order according to class and grade beginning, after the end time, can generate according to historical data Prediction window, based on start and end time, duration, it is thus achieved that prediction window quantity.

The newly-built module window banging people according to revised prediction window number data genaration each time period that arranges an order according to class and grade Member's scheduling data (utilizing 0-1 nonlinear programming to realize automatic elastic to arrange an order according to class and grade), window banging personnel arrange an order according to class and grade number According to being drawn by following object function (in order to increase motility and mobility, also for the adaptation model expanding algorithm Enclose, some fixed value of algorithm is carried out parametrization management, corresponding change object function and constraints)；

Window banging personnel are at the time parameter of window services: the staff of window is once work the most extremely Few stream time section is startmintimeslot (being abbreviated as ST, default value is 1)；Take again after rest The minimum service time continuously of business, number was middlemintimeslot (being abbreviated as MI, default value is 3), Big service time continuously, number was maxtimeslot (being abbreviated as MA, default value is 5)；On window banging personnel Noon or section are endmintimeslot (being abbreviated as EN, default value is 2) the minimum working time in the afternoon.

Hop count parametrization time shared by whole day；Time shared by the morning hop count be morningtimeslot (be abbreviated as MO, Default value is 14)；Time shared by noon, hop count is noontimeslot (being abbreviated as NO, default value is 8)；Under Time shared by noon, hop count is afternoontimeslot (being abbreviated as AF, default value is 12).

Between window banging personnel, within a cycle, net cycle time length maximum disparity parameter turns to Allowinterval (is abbreviated as AL, default value is 4).Object function is as follows:

$\left\{\begin{array}{c}{\mathrm{minz}}_1=\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{ijk}+(NO-1)\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{T}{\Σ}}{x}_{ij(MO+1)};\\ {\mathrm{minz}}_2=n.\end{array}\right.$

Wherein: T represents plan phase total natural law of arranging an order according to class and grade；N represents the sum of window banging personnel；c_jkExpression is arranged an order according to class and grade The minimum window number needed for the customer demand for service jth sky k moment is met in the plan phase；Decision variable x_ijkRepresent Certain window banging personnel every day in the duty of window banging period, when the window banging personnel day some time When being engraved in window banging, value is 1, is otherwise 0；

Description according to window shifts arrangement and analysis, and combine the definition of decision variable, the constraint bar of model Part mainly has:

$(ST-1)(1-x_{ij,k})(1-x_{ij,k+1})x_{ij,k+2}\≤\underset{n=k+3}{\overset{k+1+ST}{\Σ}}{x}_{ijn},$

$\∀i,j,k=1,2,...,MO-1-ST,MO+2,...,MO+AF-ST;$

$(ST-1)(1-x_{ij1})x_{ij2}\≤\underset{n=3}{\overset{1+ST}{\Σ}}{x}_{ijn},\∀i,j;$

$(ST-1)x_{ij1}\≤\underset{n=2}{\overset{ST}{\Σ}}{x}_{ijn},\∀i,j.$

Above formula ensure that the staff of window works ST (≤MI) the individual time period the most continuously once work；

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\≤\underset{n=k+3}{\overset{MO}{\Π}}(1-x_{ijn}),$

$\∀i,j,k=1,2,...,MO-3;$

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\≤\underset{n=k+3}{\overset{MO+1+AF}{\Π}}(1-x_{ijn}),$

$\∀i,j,k=MO+2,...,MO+AF-2;$

If above formula is had a rest after 2 time periods after ensure that staff's work of window again, then arrive that morning or ought Afternoon is the most all had a rest；

$(EN-1)x_{ijk}(1-x_{ij,k+1})x_{ij,k+2}\≤\underset{n=k+3}{\overset{k+1+EN}{\Σ}}{x}_{ijn},$

$\∀i,j,k=ST,\mathrm{...},MO-1-EN,MO+1+ST,\mathrm{...},MO+AF-EN;$

Above formula ensure that the staff of window to have a rest after 1 time period and works the most continuously once start working again EN (≤MI) the individual time period；

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\Π}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO-EN}{\Σ}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\≤(MO-2EN-k-1)\underset{l=k+2+EN}{\overset{k+1+MI}{\Σ}}{x}_{ijl},$

$\∀i,j,k=ST,...,MO-2-2EN;$

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\mathrm{\Π}}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO+AF+1-EN}{\Σ}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\≤(MO+AF-2EN-k)\underset{l=k+2+EN}{\overset{k+1+MI}{\mathrm{\Π}}}{x}_{ijl},$

$\∀i,j,k=MO+1+ST,\mathrm{...},MO+AF-2EN-1;$

Above formula ensure that the staff of window to have a rest after 1 time period to start working the most again and is not that morning Or the last working time section that afternoon is before leaving offices works MI time period the most continuously；

$\underset{d=k}{\overset{k+MA}{\Σ}}{x}_{ijd}\≤MA,\∀i,j,k=1,2,\mathrm{...},MO-MA,MO+2,\mathrm{...},MO+AF+1-MA$

Above formula ensure that window banging personnel work MA time period the most continuously once work；

$1\≤\[\underset{k=1}{\overset{MO}{\Π}}(1-x_{ijk})+(1-x_{ij,MO+1})+\underset{k=MO+2}{\overset{MO+1+AF}{\Π}}(1-x_{ijk})\]\≤2,\∀i,j$

Above formula ensure that the only work 2 of any one sky (morning, noon, afternoon) in window banging person works's day Time period or 1 time period (the darg time is less than 8 hours)；

$\frac{1}{2}(x_{ij,MO-1}+x_{ij,MO})\≤1-x_{ij,MO+1},\∀i,j$

If above formula ensure that window banging personnel, certain day noon is on duty, then he (she) within front 2 time periods on duty not In window services, so ensure that noon has the time to have lunch and has time on duty at noon before on duty；

$x_{ijMO}\≤x_{ij,MO-1},x_{ij,MO+1+AF}\≤x_{ij,MO+AF}\∀i,j$

Above formula ensure that window banging personnel's morning or afternoon can not only work 1 time period before leaving offices；

(1-x_{Ij, MO-2})x_{Ij, MO-1}≤x_{Ij, MO},

$(1-x_{ij,MO+AF-1})x_{ij,MO+AF}\≤x_{ij,MO+AF+1}\∀i,j$

Above formula ensure that window banging personnel can only work 2 time periods before leaving offices in the morning or afternoon；

$x_{ijMO}\left(\underset{k=MO+1}{\overset{MO+AF}{\Σ}}{x}_{ijk}\right)=0,\∀j,i,$

If above formula ensure window banging personnel on duty for noon, then afternoon he (she) not at window banging；

$\underset{i=1}{\overset{n}{\Σ}}{x}_{ijk}\≥c_{jk},\∀j,k$

Above formula ensure that jth sky random time section k (by historical data, utilizes the theory of statistics and queueing theory to protect Barrier client's waiting time do not surpass 15 minutes estimate window number) operate at full capacity；

$\underset{i=1}{\overset{n}{\Σ}}{x}_{ij,MO+1}=1,\∀j$

Above formula ensure that jth sky have noon and only 1 window banging personnel on duty；

$|\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{ijk}+(NO-1)\underset{j=1}{\overset{T}{\Σ}}{x}_{lj,MO+1}-\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{mjk}-(NO-1)\underset{j=1}{\overset{T}{\Σ}}{x}_{mj,MO+1}|\≤AL,\∀l,m$

Above formula ensure that when between window banging personnel, within a cycle, net cycle time length is less than AL Between section (embody fair principle).

History contemporaneous data management module in, the history business datum same period include business item, the time of drawing a bill, Waiting time, deadline, accept window and working window cadre.

In window number generation module, client average latency W_qLess than 15 minutes, i.e. W_q≤ 0.25 hour.

User and staff's benefits module, for record window staff's personal information, window banging personnel People's information includes name, job number and telephone number.

Algorithm parameter management module, for recording the various parameters needed for prediction window number and algorithm of arranging an order according to class and grade Enter and dynamically manage, including traffic handling time, business weight, queue length, client's latency tolerant Maximum disparity between the upper limit, employee's High-effective Service duration, period maximal workload and minimum workload requires, Participation is arranged an order according to class and grade time shared by quantity of cadres, noon hop count and maximum motor-driven staff number every day.

Segment management module when arranging an order according to class and grade, specifically the starting and tie of period of when prediction window number, every day being classified The bundle time carries out typing and management.

Arrange an order according to class and grade management module, measure query statistic, data backup, reduction, system for the realization person works that arranges an order according to class and grade User and rights management.

The example of model solves

Obviously, above-mentioned model is a nonlinear integer programming model, and integer programming is a NPC problem, Cannot ensure in polynomial time, try to achieve optimal solution, it is true that we utilize Lingo program calculation one Week does not the most obtain optimal solution or near-optimal solution. and we select c language that variable is carried out dimension-reduction treatment, warp herein Substantial amounts of data test, result shows to be no more than on common computer to calculate optimal solution in two minutes.

After we carry out statistical disposition to the historical data of Yunhe County of bureau of local taxation, Lishui, Zhejiang Province, obtain pre- Survey in January, 2016 following data.

	Taxpayer reaches to service the number in the Room	Taxpayer reaches to service the number in the Room
			Time period	2016.1.4-1.8 average number every day	2016.1.11-1.15 average number every day
8:31-9:30	13	10
			9:31-10:30	12	10
10:31-11:30	10	4
			11:31-12:00	2	1
12:01-14:00	1	1
			14:01-15:00	4	2
15:01-16:00	12	12
			16:01-17:00	9	8

The cadre that pays taxes of window averagely handles a business 5.6 minutes, and taxpayer waits that average time is less than 15 minutes, go out, by the M/M/C model pre-estimating of queueing theory, the window number that January, the last fortnight day part was opened.

The 0-1 nonlinear programming model set up by us and software, calculate the cadre that pays taxes elastic on duty Shift report.

Employee's net cycle time

1	2	3	4	5
					182	183	182	182	181

Numeral 1, represents the window banging personnel state at window banging, numeral 0, represents window banging personnel Arraying data on backstage or do other thing, window banging personnel 1,2,3,4,5 (are within a cycle Saving space, we take two weeks is a cycle) basically identical in the time of window banging.Can by example Seeing, the present invention proposes the solution of elastic services and more manually arranges an order according to class and grade and have the advantage that

1. the constraints of this example window personnel flexible scheduling is more than 2000, and decision variable is more than 2000 Individual, if manually arranging an order according to class and grade, solve optimal solution extremely difficult.

2. can effectively control the custom queueing waiting time by scientific algorithm with arranging an order according to class and grade.

3. elastic service shift scheduling model processed can dynamically change according to demand adjust number that window goes to work and Working continuously and the time having a rest is adjusted according to local demand.

Illustrate as a example by Local Tax Bureau of Lishui Yunhe County now.After elastic service system is carried out, receive Tax people, a linear window are paid taxes work cadre, and window management layer three aspect achieves win-win.Service person-time, industry The business amount of handling increases on year-on-year basis.Tax handling service hall, the phenomenons such as personnel are empty are greatly reduced, and work efficiency is averagely done Reason time, taxpayer's satisfaction have lifting.After elastic service pilot processed, taxpayer's waiting time is the longest not More than 15 minutes, the satisfaction of service was improved by taxpayer.Tax handling service hall staff window banging duration Be greatly shortened, shortened to from the 18.5 ten thousand of the same period last year hours more than 80,000 hour, actual accept the business hours with The ratio of total window line duration brings up to 44.45% from 22.67%, and other data see table.

Note: the last week :-21 days on the 17th November；One week after :-28 days on the 24th November.

The foregoing is only the preferred embodiment of the present invention, protection scope of the present invention is not limited in above-mentioned Embodiment, every technical scheme belonging to the principle of the invention belongs to protection scope of the present invention.For this For the technical staff in field, the some improvement carried out on the premise of without departing from the principle of the present invention, these Improve and also should be regarded as protection scope of the present invention.

Claims (9)

1. one kind based on parameterized elastic services scheduling method, it is characterised in that include walking as follows Rapid:

According to the window number that the history business datum same period prediction subsequent work time period is open, subsequent work time The minimum window number of Duan Kaifang is drawn by equation below:

$\begin{array}{c}{W}_q=\frac{L_q}{\mathrm{\λ}}=\frac{{\left(c\mathrm{\ρ}\right)}^c\mathrm{\ρ}}{c!{(1-\mathrm{\ρ})}^2}{P}_0=\\ \frac{{\left(c\mathrm{\ρ}\right)}^c\mathrm{\ρ}}{c!{(1-\mathrm{\ρ})}^2}{\[\underset{k=0}{\overset{c-1}{\Σ}}\frac{{\left(c\mathrm{\ρ}\right)}^k}{k!}+\frac{{\left(c\mathrm{\ρ}\right)}^c}{c!(1-\mathrm{\ρ})}\]}^{-1}.\end{array}---\left(1\right)$

Wherein, W_qRepresent client's average latency；C is open minimum window number needed for representing the service Room；Represent service intensity；λ represents the average client's number arriving the service Room per hour；μ represents average every Client's number that one window can service per hour；

t₀Representing the average latency upper limit that client can tolerate, for formula (1), calculation procedure is as follows:

Step 1: order

Step 2: calculate W according to formula (1)_qIf, W_q≤t₀, then terminating algorithm, c now is required Minimum window number；If W_q＞ t₀Then go to step 3；

Step 3: make c=c+1, goes to step 2；

According to window banging personnel's scheduling data of revised prediction window number data genaration each time period, Window banging personnel's scheduling data is drawn by following object function；

Window banging personnel are at the time parameter of window services: the staff of window is once work the most extremely Few stream time section is startmintimeslot, is abbreviated as ST, and default value is 1；Take again after rest The minimum service time continuously of business, number was middlemintimeslot, was abbreviated as MI, and default value is 3, maximum Service time, number was maxtimeslot continuously, was abbreviated as MA, and default value is 5；Window banging personnel's morning or Section is endmintimeslot the minimum working time in the afternoon, is abbreviated as EN, and default value is 2；

Hop count parametrization time shared by whole day；Time shared by the morning, hop count is morningtimeslot, is abbreviated as MO, Default value is 14；Time shared by noon, hop count is noontimeslot, is abbreviated as NO, and default value is 8；Afternoon Time shared, hop count is afternoontimeslot, is abbreviated as AF, and default value is 12；

Between window banging personnel, within a cycle, net cycle time length maximum disparity parameter turns to Allowinterval, is abbreviated as AL, and default value is 4；

Object function is:

$\left\{\begin{array}{c}{\mathrm{minz}}_1=\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{T}{\Σ}}\underset{k=1}{\overset{MO+AF+1}{\Σ}}{x}_{ijk}+(NO-1)\underset{i=1}{\overset{n}{\Σ}}\underset{j=1}{\overset{T}{\Σ}}{x}_{ij(MO+1)};\\ {\mathrm{minz}}_2=n.\end{array}\right.$

Wherein: T represents plan phase total natural law of arranging an order according to class and grade；N represents the sum of window banging personnel；c_jkExpression is arranged an order according to class and grade The minimum window number needed for the customer demand for service jth sky k moment is met in the plan phase；Decision variable x_ijkRepresent Certain window banging personnel every day in the duty of window banging period, when the window banging personnel day some time When being engraved in window banging, value is 1, is otherwise 0；

Constraints:

$(ST-1)(1-x_{ij,k})(1-x_{ij,k+1})x_{ij,k+2}\≤\underset{n=k+3}{\overset{k+1+ST}{\Σ}}{x}_{ijn},$

$\∀i,j,k=1,2,...,MO-1-ST,MO+2,...,MO+AF-ST;$

$(ST-1)(1-x_{ij1})x_{ij2}\≤\underset{n=3}{\overset{1+ST}{\Σ}}{x}_{ijn},\∀i,j;$

$(ST-1)x_{ij1}\≤\underset{n=2}{\overset{ST}{\Σ}}{x}_{ijn},\∀i,j;$

Above formula ensure that the staff of window works ST time period the most continuously once work, ST≤MI；

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\≤\underset{n=k+3}{\overset{MO}{\Π}}(1-x_{ijn}),$

$\∀i,j,k=1,2,...,MO-3;$

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\≤\underset{n=k+3}{\overset{MO+1+AF}{\Π}}(1-x_{ijn}),$

$\∀i,j,k=MO+2,...,MO+AF-2;$

If above formula is had a rest after 2 time periods after having demonstrate,proved staff's work of window again, then arrive that morning or ought Afternoon is the most all had a rest；

$(EN-1)x_{ijk}(1-x_{ij,k+1})x_{ij,k+2}\≤\underset{n=k+3}{\overset{k+1+EN}{\Σ}}{x}_{ijn},$

$\∀i,j,k=ST,...,MO-1-EN,MO+1+ST,...,MO+AF-EN;$

Above formula ensure that the staff of window to have a rest after 1 time period and works the most continuously once start working again EN time period, EN≤MI；

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\Π}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO-EN}{\Σ}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\≤(MO-2EN-k-1)\underset{l=k+2+EN}{\overset{k+1+MI}{\Π}}{x}_{ijl},$

$\∀i,j,k=ST,...,MO-2-2EN;$

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\Π}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO+AF+1-EN}{\Σ}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\≤(MO+AF-2EN-k)\underset{l=k+2+EN}{\overset{k+1+MI}{\Π}}{x}_{ijl},$

$\∀i,j,k=MO+1+ST,...,MO+AF-2EN-1;$

Above formula ensure that the staff of window to have a rest after 1 time period to start working the most again and is not when the sky Noon or afternoon, last working time section before leaving offices worked MI time period the most continuously；

$\underset{d=k}{\overset{k+MA}{\Σ}}{x}_{ijd}\≤MA,\∀i,j,k=1,2,...,MO-MA,MO+2,...,MO+AF+1-MA$

Above formula ensure that window banging personnel work MA time period the most continuously once work；

$1<\&lsqb;\underset{k=1}{\overset{MO}{\&Pi;}}(1-x_{ijk})+(1-x_{ij,MO+1})+\underset{k=MO+2}{\overset{MO+1+AF}{\&Pi;}}(1-x_{ijk})\&rsqb;\&le;2,\&ForAll;i,j$

Above formula ensure that any one sky in window banging person works's day, the i.e. morning, noon, afternoon, only work 2 Individual time period or 1 time period, the darg time is less than 8 hours；

$\frac{1}{2}(x_{ij,MO-1}+x_{ij,MO})\&le;1-x_{ij,MO+1},\&ForAll;i,j$

If above formula ensure that window banging personnel, certain day noon is on duty, then he (she) within front 2 time periods on duty not In window services, so ensure that noon has the time to have lunch and has time on duty at noon before on duty；

$x_{ijMO}\&le;x_{ij,MO-1},x_{ij,MO+1+AF}\&le;x_{ij,MO+AF}\&ForAll;i,j$

Above formula ensure that window banging personnel's morning or afternoon can not only work 1 time period before leaving offices；

(1-x_{Ij, MO-2})x_{Ij, MO-1}≤x_{Ij, MO},

$(1-x_{ij,MO+AF-1})x_{ij,MO+AF}\&le;x_{ij,MO+AF+1}\&ForAll;i,j$

Above formula ensure that window banging personnel can only work 2 time periods before leaving offices in the morning or afternoon；

$x_{ijMO}\left(\underset{k=MO+1}{\overset{MO+AF}{\&Sigma;}}{x}_{ijk}\right)=0,\&ForAll;j,i,$

If above formula ensure window banging personnel on duty for noon, then afternoon he (she) not at window banging；

$\underset{i=1}{\overset{n}{\&Sigma;}}{x}_{ijk}\&GreaterEqual;c_{jk},\&ForAll;j,k$

Above formula ensure that jth sky random time section k is operated at full capacity；

$\underset{i=1}{\overset{n}{\&Sigma;}}{x}_{ij,MO+1}=1,\&ForAll;j$

Above formula ensure that jth sky have noon and only 1 window banging personnel on duty；

$|\underset{j=1}{\overset{T}{\&Sigma;}}\underset{k=1}{\overset{MO+AF+1}{\&Sigma;}}{x}_{ljk}+(NO-1)\underset{j=1}{\overset{T}{\&Sigma;}}{x}_{lj,MO+1}-\underset{j=1}{\overset{T}{\&Sigma;}}\underset{k=1}{\overset{MO+AF+1}{\&Sigma;}}{x}_{mjk}-(NO-1)\underset{j=1}{\overset{T}{\&Sigma;}}{x}_{mj,MO+1}|\&le;AL,\&ForAll;l,m$

Above formula ensures that between window banging personnel, net cycle time length is less than AL time period within a cycle.

The most according to claim 1 based on parameterized elastic services scheduling method, its feature It is: the history business datum same period includes business item, time of drawing a bill, waiting time, deadline, is subject to Reason window and window banging personnel.

The most according to claim 1 based on parameterized elastic services scheduling method, its feature It is: client average latency W_qLess than 15 minutes, i.e. W_q≤ 0.25 hour.

4. one kind based on parameterized elastic services Workforce Management, it is characterised in that including:

History contemporaneous data management module, for the record of historical data and management:

Window number generation module, according to the window that the history business datum same period prediction subsequent work time period is open Number, the minimum window number that the subsequent work time period opens is drawn by equation below:

$\begin{array}{c}{W}_q=\frac{L_q}{\mathrm{\&lambda;}}=\frac{{\left(c\mathrm{\&rho;}\right)}^c\mathrm{\&rho;}}{c!{(1-\mathrm{\&rho;})}^2}{P}_0=\\ \frac{{\left(c\mathrm{\&rho;}\right)}^c\mathrm{\&rho;}}{c!{(1-\mathrm{\&rho;})}^2}{\&lsqb;\underset{k=0}{\overset{c-1}{\&Sigma;}}\frac{{\left(c\mathrm{\&rho;}\right)}^k}{k!}+\frac{{\left(c\mathrm{\&rho;}\right)}^c}{c!(1-\mathrm{\&rho;})}\&rsqb;}^{-1}.\end{array}---\left(1\right)$

Wherein, W_qRepresent client's average latency；C is open minimum window number needed for representing the service Room；Represent service intensity；λ represents the average client's number arriving the service Room per hour；μ represents average every Client's number that one window can service per hour；

t₀Representing the average latency upper limit that client can tolerate, for formula (1), calculation procedure is as follows:

Step 1: order

Step 2: calculate W according to formula (1)_qIf, W_q≤t₀, then terminating algorithm, c now is required Minimum window number；If W_q＞ t₀Then go to step 3；

Step 3: make c=c+1, goes to step 2；

Newly-built module of arranging an order according to class and grade, according to the window banging of revised prediction window number data genaration each time period Personnel's scheduling data, window banging personnel's scheduling data is drawn by following object function；

Window banging personnel are at the time parameter of window services: the staff of window is once work the most extremely Few stream time section is startmintimeslot, is abbreviated as ST, and default value is 1；Take again after rest The minimum service time continuously of business, number was middlemintimeslot, was abbreviated as MI, and default value is 3, maximum Service time, number was maxtimeslot continuously, was abbreviated as MA, and default value is 5；Window banging personnel's morning or Section is endmintimeslot the minimum working time in the afternoon, is abbreviated as EN, and default value is 2；

Hop count parametrization time shared by whole day；Time shared by the morning, hop count is morningtimeslot, is abbreviated as MO, Default value is 14；Time shared by noon, hop count is noontimeslot, is abbreviated as NO, and default value is 8；Afternoon Time shared, hop count is afternoontimeslot, is abbreviated as AF, and default value is 12；

Between window banging personnel, within a cycle, net cycle time length maximum disparity parameter turns to Allowinterval, is abbreviated as AL, and default value is 4；

Object function is:

$\left\{\begin{array}{c}{\mathrm{minz}}_1=\underset{i=1}{\overset{n}{\&Sigma;}}\underset{j=1}{\overset{T}{\&Sigma;}}\underset{k=1}{\overset{MO+AF+1}{\&Sigma;}}{x}_{ijk}+(NO-1)\underset{i=1}{\overset{n}{\&Sigma;}}\underset{j=1}{\overset{T}{\&Sigma;}}{x}_{ij(MO+1)};\\ {\mathrm{minz}}_2=n.\end{array}\right.$

Wherein: T represents plan phase total natural law of arranging an order according to class and grade；N represents the sum of window banging personnel；c_jkExpression is arranged an order according to class and grade The minimum window number needed for the customer demand for service jth sky k moment is met in the plan phase；Decision variable x_ijkRepresent Certain window banging personnel every day in the duty of window banging period, when the window banging personnel day some time When being engraved in window banging, value is 1, is otherwise 0；

Constraints:

$(ST-1)(1-x_{ij,k})(1-x_{ij,k+1})x_{ij,k+2}\&le;\underset{n=k+3}{\overset{k+1+ST}{\&Sigma;}}{x}_{ijn},$

$\&ForAll;i,j,k=1,2,...,MO-1-ST,MO+2,...,MO+AF-ST;$

$(ST-1)(1-x_{ij1})x_{ij2}\&le;\underset{n=3}{\overset{1+ST}{\&Sigma;}}{x}_{ijn},\&ForAll;i,j;$

$(ST-1)x_{ij1}\&le;\underset{n=2}{\overset{ST}{\&Sigma;}}{x}_{ijn},\&ForAll;i,j;$

Above formula ensure that the staff of window works ST time period the most continuously once work, ST≤MI；

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\&le;\underset{n=k+3}{\overset{MO}{\&Pi;}}(1-x_{ijn}),$

$\&ForAll;i,j,k=1,2,...,MO-3;$

$x_{ijk}(1-x_{ij,k+1})(1-x_{ij,k+2})\&le;\underset{n=k+3}{\overset{MO+1+AF}{\&Pi;}}(1-x_{ijn}),$

$\&ForAll;i,j,k=MO+2,...,MO+AF-2;$

If above formula is had a rest after 2 time periods after ensure that staff's work of window again, then arrive that morning or ought Afternoon is the most all had a rest；

$(EN-1)x_{ijk}(1-x_{ij,k+1})x_{ij,k+2}\&le;\underset{n=k+3}{\overset{k+1+EN}{\&Sigma;}}{x}_{ijn},$

$\&ForAll;i,j,k=ST,...,MO-1-EN,MO+1+ST,...,MO+AF-EN;$

Above formula ensure that the staff of window to have a rest after 1 time period and works the most continuously once start working again EN time period, EN≤MI；

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\&Pi;}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO-EN}{\&Sigma;}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\&le;(MO-2EN-k-1)\underset{l=k+2+EN}{\overset{k+1+MI}{\&Pi;}}{x}_{ijl},$

$\&ForAll;i,j,k=ST,...,MO-2-2EN;$

$x_{ijk}(1-x_{ij,k+1})\underset{n=k+2}{\overset{k+1+EN}{\&Pi;}}{x}_{ijn}\left(\underset{m=k+2+EN}{\overset{MO+AF+1-EN}{\&Sigma;}}\right(1-x_{ijm}\left)x_{ij,m+1}\right)\&le;(MO+AF-2EN-k)\underset{l=k+2+EN}{\overset{k+1+MI}{\&Pi;}}{x}_{ijl},$

$\&ForAll;i,j,k=MO+1+ST,...,MO+AF-2EN-1;$

Above formula demonstrate,proved the staff of window have a rest after 1 time period start working the most again and be not that morning or Afternoon, last working time section before leaving offices worked MI time period the most continuously；

$\underset{d=k}{\overset{k+MA}{\&Sigma;}}{x}_{ijd}\&le;MA,\&ForAll;i,j,k=1,2,...,MO-MA,MO+2,...,MO+AF+1-MA$

Above formula ensure that window banging personnel work MA time period the most continuously once work；

$1<\&lsqb;\underset{k=1}{\overset{MO}{\&Pi;}}(1-x_{ijk})+(1-x_{ij,MO+1})+\underset{k=MO+2}{\overset{MO+1+AF}{\&Pi;}}(1-x_{ijk})\&rsqb;\&le;2,\&ForAll;i,j$

Above formula ensure that any one sky in window banging person works's day, the i.e. morning, noon, afternoon, only work 2 Individual time period or 1 time period, the darg time is less than 8 hours；

$\frac{1}{2}(x_{ij,MO-1}+x_{ij,MO})\&le;1-x_{ij,MO+1},\&ForAll;i,j$

If above formula ensure that window banging personnel, certain day noon is on duty, then he (she) within front 2 time periods on duty not In window services, so ensure that noon has the time to have lunch and has time on duty at noon before on duty；

$x_{ijMO}\&le;x_{ij,MO-1},x_{ij,MO+1+AF}\&le;x_{ij,MO+AF}\&ForAll;i,j$

Above formula ensure that window banging personnel's morning or afternoon can not only work 1 time period before leaving offices；

(1-x_{Ij, MO-2})x_{Ij, MO-1}≤x_{Ij, MO},

$(1-x_{ij,MO+AF-1})x_{ij,MO+AF}\&le;x_{ij,MO+AF+1}\&ForAll;i,j$

Above formula ensure that window banging personnel can only work 2 time periods before leaving offices in the morning or afternoon；

$x_{ijMO}\left(\underset{k=MO+1}{\overset{MO+AF}{\&Sigma;}}{x}_{ijk}\right)=0,\&ForAll;j,i,$

If above formula ensure window banging personnel on duty for noon, then afternoon he (she) not at window banging；

$\underset{i=1}{\overset{n}{\&Sigma;}}{x}_{ijk}\&GreaterEqual;c_{jk},\&ForAll;j,k$

Above formula ensure that jth sky random time section k is operated at full capacity；

$\underset{i=1}{\overset{n}{\&Sigma;}}{x}_{ij,MO+1}=1,\&ForAll;j$

Above formula ensure that jth sky have noon and only 1 window banging personnel on duty；

$|\underset{j=1}{\overset{T}{\&Sigma;}}\underset{k=1}{\overset{MO+AF+1}{\&Sigma;}}{x}_{ljk}+(NO-1)\underset{j=1}{\overset{T}{\&Sigma;}}{x}_{lj,MO+1}-\underset{j=1}{\overset{T}{\&Sigma;}}\underset{k=1}{\overset{MO+AF+1}{\&Sigma;}}{x}_{mjk}-(NO-1)\underset{j=1}{\overset{T}{\&Sigma;}}{x}_{mj,MO+1}|\&le;AL,\&ForAll;l,m$

Above formula ensures that between window banging personnel, net cycle time length is less than AL time period within a cycle.

The most according to claim 4 based on parameterized elastic services Workforce Management, its feature Be: history contemporaneous data management module in, the history business datum same period include business item, the time of drawing a bill, Waiting time, deadline, accept window and working window cadre.

The most according to claim 4 based on parameterized elastic services Workforce Management, its feature It is: in window number generation module, client average latency W_qLess than 15 minutes, i.e. W_q≤ 0.25 is little Time.

The most according to claim 4 based on parameterized elastic services Workforce Management, its feature It is: include user and staff's benefits module, for record window staff's personal information, window banging Personnel's personal information includes name, job number and telephone number.

The most according to claim 4 based on parameterized elastic services Workforce Management, its feature It is: include that algorithm parameter manages module, for the various parameters needed for prediction window number and algorithm of arranging an order according to class and grade Carry out typing and dynamically management, including traffic handling time, business weight, queue length, client's tolerance etc. Treat that the maximum between time upper limit, employee's High-effective Service duration, period maximal workload and minimum workload is poor Hop count and maximum motor-driven staff number every day in time requiring, participate in arrange an order according to class and grade shared by quantity of cadres, noon.

The most according to claim 4 based on parameterized elastic services Workforce Management, its feature It is: segment management module when including arranging an order according to class and grade, specifically opening of period of when prediction window number, every day being classified Begin and the end time carries out typing and management.