CN116128558A - Intelligent scheduling time sequence prediction method, system, equipment and storage medium - Google Patents
Intelligent scheduling time sequence prediction method, system, equipment and storage medium Download PDFInfo
- Publication number
- CN116128558A CN116128558A CN202211521463.3A CN202211521463A CN116128558A CN 116128558 A CN116128558 A CN 116128558A CN 202211521463 A CN202211521463 A CN 202211521463A CN 116128558 A CN116128558 A CN 116128558A
- Authority
- CN
- China
- Prior art keywords
- result
- matrix
- scheduling
- historical
- traffic data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 230000000737 periodic effect Effects 0.000 claims abstract description 103
- 230000008859 change Effects 0.000 claims abstract description 95
- 238000012795 verification Methods 0.000 claims abstract description 80
- 238000004458 analytical method Methods 0.000 claims abstract description 53
- 238000012549 training Methods 0.000 claims abstract description 44
- 239000011159 matrix material Substances 0.000 claims description 169
- 238000012545 processing Methods 0.000 claims description 60
- 238000004364 calculation method Methods 0.000 claims description 22
- 230000006870 function Effects 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 5
- 230000001360 synchronised effect Effects 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06316—Sequencing of tasks or work
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/109—Time management, e.g. calendars, reminders, meetings or time accounting
- G06Q10/1093—Calendar-based scheduling for persons or groups
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The embodiment of the invention discloses an intelligent scheduling time sequence prediction method, a system, equipment and a storage medium, wherein, firstly, historical telephone traffic data of a call center is obtained according to a preset period; then, based on time sequence distribution characteristics of historical telephone traffic data, carrying out characteristic analysis on the historical telephone traffic data to obtain corresponding data change characteristics; then, the data change characteristics are self-attentively encoded by utilizing the periodic change rule of the data change characteristics, so that the corresponding periodic characteristics are obtained; training to obtain a human demand prediction model based on historical telephone traffic data and corresponding periodic characteristics; and generating a first scheduling result by using a human demand prediction result predicted by the human demand prediction model, performing verification analysis on the first scheduling result, and taking the verified first scheduling result as a second scheduling result. The embodiment of the invention effectively improves the practicability of the scheduling result, realizes the timely update of the scheduling result, and simultaneously avoids the problem of overlarge working hour difference among staff.
Description
Technical Field
The embodiment of the invention relates to the technical field of communication, in particular to an intelligent scheduling time sequence prediction method, system, equipment and storage medium.
Background
With the rapid development of related platforms and network services of big data and artificial intelligence at present, the scale of a call center is rapidly increased, the service structure is increasingly complex, and the reasonable configuration of human resources of a call center system is required to be realized through an efficient and economical scheduling scheme, so that the minimization of operation cost and the maximization of profit are realized, and the service level and the service quality of clients are effectively ensured.
The existing scheduling technology needs to preset appointed parameters, such as an average value of historical telephone traffic, performs scheduling according to the preset appointed parameters, does not utilize the internal relation and the change rule among the historical telephone traffic data, and has low practicability of scheduling results; meanwhile, the collection period of the historical telephone traffic data is longer, and when special change occurs, the scheduling result cannot be updated in time; the scheduling process of the existing scheduling technology needs manual intervention, has large human influence factors, has the problem of large man-hour difference among different staff, and cannot reasonably distribute human resources.
Disclosure of Invention
Therefore, the embodiment of the invention provides an intelligent scheduling time sequence prediction method, system, equipment and storage medium, which are used for solving the problems of low practicality, poor updating capability and large staff man-hour difference of the scheduling result of the existing scheduling technology.
In order to achieve the above object, the embodiment of the present invention provides the following technical solutions:
according to a first aspect of an embodiment of the present invention, there is provided an intelligent scheduling timing prediction method, the method including:
according to a preset period, acquiring historical telephone traffic data of a call center, wherein the historical telephone traffic data comprises a historical call-in request quantity, a historical call completing rate, a historical personnel attendance rate and a historical personnel utilization rate;
based on the time sequence distribution characteristics of the historical telephone traffic data, performing first characteristic analysis processing on the historical telephone traffic data to obtain first data change characteristics; performing second characteristic analysis processing on the first data change characteristic according to the data change rule of the first data change characteristic to obtain a second data change characteristic;
according to the periodic variation rule of the second data variation characteristic, performing first self-attention coding processing on the second data variation characteristic to obtain a corresponding periodic variation characteristic; performing second self-attention coding processing on the periodic variation characteristics to obtain periodic fluctuation characteristics;
Performing iterative training on the human demand prediction model to be trained based on the historical telephone traffic data, the periodic variation characteristics and the periodic fluctuation characteristics to obtain a trained human demand prediction model;
according to the historical telephone traffic data, scheduling telephone traffic data of a period to be scheduled is obtained through prediction, the scheduling telephone traffic data is input into the trained human demand prediction model, a human demand prediction result is obtained, and a first scheduling result is generated according to the human demand prediction result;
and carrying out verification analysis processing on the first scheduling result to obtain a verification analysis result, and taking the first scheduling result as a second scheduling result if the verification analysis result is qualified.
Further, based on the time sequence distribution characteristic of the historical traffic data, performing a first characteristic analysis processing on the historical traffic data to obtain a first data change characteristic, including:
according to the historical telephone traffic data, an input data matrix alpha is obtained, and each column of the input data matrix alpha comprises the historical incoming call request quantity, the historical call completing rate, the historical personnel attendance rate and the historical personnel utilization rate in one preset period;
Based on the time sequence distribution characteristics of the historical telephone traffic data, carrying out probability density operation on the input data matrix alpha to obtain a probability density matrix A corresponding to the input data matrix alpha, wherein the calculation formula of the probability density matrix A is as follows:
A＝{A i }
wherein A is i An i-th element representing the probability density matrix a; alpha i An ith column representing the input data matrix α;
is formed by alpha i A column matrix formed by any value in the value range; />
Representing the Euclidean distance between the maximum value and the average value in the input data matrix alpha; Δα represents an increment of the input data matrix α; i is an integer greater than or equal to zero and less than the total number of historical traffic data periods.
Further, according to the data change rule of the first data change feature, performing a second feature analysis on the first data change feature to obtain a second data change feature, including:
performing integral operation on the probability density matrix A to obtain a parameter distribution matrix B, wherein the calculation formula of the parameter distribution matrix B is as follows:
B＝{B i }
wherein t is a preset period parameter.
Further, according to the periodic variation rule of the second data variation characteristic, performing first self-attention coding processing on the second data variation characteristic to obtain a corresponding periodic variation characteristic; and performing second self-attention coding processing on the periodic variation characteristic to obtain a periodic fluctuation characteristic, wherein the method comprises the following steps of:
Presetting a sequence number weight parameter column matrix Q, a key value weight parameter column matrix K and a numerical weight parameter column matrix V;
and performing first self-attention encoding processing on the parameter distribution matrix B by using the sequence number weight parameter column matrix Q, the key value weight parameter column matrix K and the numerical value weight parameter column matrix V to obtain a periodic variation characteristic matrix C, wherein the calculation formula of the periodic variation characteristic matrix C is as follows:
C＝{C i }
wherein C is i An ith element representing the periodically varying feature matrix C; k (K) T Representing a transposed matrix of the key value weight parameter column matrix K; d, d k The method comprises the steps of presetting a key value weight parameter; e is a natural constant;
obtaining a periodic column coefficient matrix W according to the sequence number weight parameter column matrix Q, the key value weight parameter column matrix K and the numerical value weight parameter column matrix V o ；
Using the periodic column coefficient matrix W o And performing second self-attention coding processing on the periodic variation characteristic matrix C to obtain a periodic fluctuation quantity matrix D.
Further, based on the historical traffic data, the periodic variation characteristic and the periodic fluctuation characteristic, performing iterative training on a human demand prediction model to be trained to obtain a trained human demand prediction model, including:
Inputting the input data matrix alpha, the periodic characteristic change matrix C and the periodic fluctuation quantity matrix D into a prediction model phi to be trained, and calculating to obtain a human demand training result x, wherein the calculation formula of the human demand training result x is as follows:
x＝Φ(α，C，D)＝θ[{max(0，wα+b)}+{C，D}]
wherein max is a preset full link layer activation function; w is a coefficient feature matrix; b is a preset induced deviation residual error; θ is a preset model parameter;
according to the human demand training result x, the coefficient feature matrix w is adjusted until the optimal human demand training result x is predicted by fitting the prediction model phi;
and obtaining the human demand prediction model after training by using the coefficient characteristic matrix w after adjustment.
Further, according to the historical traffic data, scheduling traffic data of a period to be scheduled is obtained through prediction, the scheduling traffic data is input into the trained human demand prediction model, a human demand prediction result is obtained, and according to the human demand prediction result, a first scheduling result is generated, and the method comprises the following steps:
according to historical synchronous traffic data and recent traffic data in the historical traffic data, predicting to obtain scheduling traffic data of a period to be scheduled, wherein the scheduling traffic data comprises at least one predicted incoming call request quantity, predicted call completing rate, predicted personnel attendance rate and predicted personnel utilization rate of a preset period;
Inputting the scheduling telephone traffic data into the trained human demand prediction model, and outputting a corresponding human demand prediction result y;
matching the human demand prediction result y corresponding to each preset period with a preset working schedule to obtain a human demand list of each working time period of the period to be scheduled;
aiming at each shift group, according to the manpower demand list of each working time period, the method aims at the staffGroup allocation is carried out to obtain the number n of group workers of the shift group in each working period j The number n of team workers in each working period j The calculation formula of (2) is as follows:
wherein k is the total preset cycle number of the period to be scheduled; h is a j The staff utilization rate is preset; y is j The human demand prediction results of the corresponding working time periods in the human demand list of each working time period are obtained; j is an integer greater than or equal to 1 and less than or equal to k;
team working number n using each working time period j The manual demand list of each working time period is used for obtaining the shift and change results of each working time period;
generating a personal scheduling result of each staff according to the group rotation result of each working time period;
and taking the shift results of each working time period group and the personal shift result as the first shift result.
Further, performing verification analysis processing on the first shift result to obtain a verification analysis result, and if the verification analysis result is qualified, taking the first shift result as a second shift result, including:
aiming at the group staff number of each shift group, calculating the difference value between the group staff number of the current group and the group staff numbers of the rest groups to obtain a group staff number difference value;
judging whether the difference value of the number of the teams is larger than a preset number difference value;
if the difference value of the number of the teams is larger than the preset difference value of the number of the teams, the verification result of the number of the teams is that the teams do not pass;
if the difference value of the number of the teams is not larger than the preset difference value of the number of the teams, the verification result of the number of the teams is passed;
counting to obtain the total scheduling working time of all staff by using all the personal scheduling results;
obtaining the average working time of the scheduling according to the total scheduling working time and the total number of workers;
aiming at each staff, the corresponding personal scheduling result is utilized to obtain the scheduling personal working time;
subtracting the working time of the shift individual from the average working time of the shift, and processing the absolute value of the operation result to obtain the absolute value of the working time difference;
Judging whether the absolute value of the working time difference value is smaller than or equal to a preset working time threshold value;
if the absolute value of the working time difference is smaller than or equal to the preset working time threshold, the staff working time length verification result is passed;
if the absolute value of the working time difference is larger than the preset working time threshold, the staff working time length verification result is that the staff working time length verification result is not passed;
judging whether the team number verification result and the employee working time verification result are both passed;
if the team number verification result and the staff work time verification result are both passed, the first scheduling result is used as the second scheduling result;
if the verification result of the number of the teams and/or the verification result of the working time of the staff is not passed, discarding the first scheduling result and performing scheduling again.
According to a second aspect of an embodiment of the present invention, there is provided an intelligent shift schedule prediction system, the system including:
the telephone traffic data acquisition module is used for acquiring historical telephone traffic data of the call center according to a preset period, wherein the historical telephone traffic data comprises a historical incoming call request quantity, a historical call completing rate, a historical personnel attendance rate and a historical personnel utilization rate;
The first data processing module is used for carrying out first characteristic analysis processing on the historical telephone traffic data based on the time sequence distribution characteristics of the historical telephone traffic data to obtain first data change characteristics; performing second characteristic analysis processing on the first data change characteristic according to the data change rule of the first data change characteristic to obtain a second data change characteristic;
the second data processing module is used for carrying out first self-attention coding processing on the second data change characteristics according to the periodic change rule of the second data change characteristics to obtain corresponding periodic change characteristics; performing second self-attention coding processing on the periodic variation characteristics to obtain periodic fluctuation characteristics;
the prediction model training module is used for carrying out iterative training on the human demand prediction model to be trained based on the historical telephone traffic data, the periodic variation characteristics and the periodic fluctuation characteristics to obtain a trained human demand prediction model;
the scheduling result generation module is used for predicting and obtaining scheduling traffic data of a period to be scheduled according to the historical traffic data, inputting the scheduling traffic data into the trained human demand prediction model to obtain a human demand prediction result, and generating a first scheduling result according to the human demand prediction result;
And the shift result verification module is used for carrying out verification analysis processing on the first shift result to obtain a verification analysis result, and if the verification analysis result is qualified, the first shift result is used as a second shift result.
Further, based on the time sequence distribution characteristic of the historical traffic data, performing a first characteristic analysis processing on the historical traffic data to obtain a first data change characteristic, including:
according to the historical telephone traffic data, an input data matrix alpha is obtained, and each column of the input data matrix alpha comprises the historical incoming call request quantity, the historical call completing rate, the historical personnel attendance rate and the historical personnel utilization rate in one preset period;
based on the time sequence distribution characteristics of the historical telephone traffic data, carrying out probability density operation on the input data matrix alpha to obtain a probability density matrix A corresponding to the input data matrix alpha, wherein the calculation formula of the probability density matrix A is as follows:
wherein A is i An i-th element representing the probability density matrix a; alpha i An ith column representing the input data matrix α;
is formed by alpha i A column matrix formed by any value in the value range; />
Representing the Euclidean distance between the maximum value and the average value in the input data matrix alpha; Δα represents an increment of the input data matrix α; i is an integer greater than or equal to zero and less than the total number of historical traffic data periods.
Further, according to the data change rule of the first data change feature, performing a second feature analysis on the first data change feature to obtain a second data change feature, including:
performing integral operation on the probability density matrix A to obtain a parameter distribution matrix B, wherein the calculation formula of the parameter distribution matrix B is as follows:
B＝{B i }
wherein t is a preset period parameter.
Further, according to the periodic variation rule of the second data variation characteristic, performing first self-attention coding processing on the second data variation characteristic to obtain a corresponding periodic variation characteristic; and performing second self-attention coding processing on the periodic variation characteristic to obtain a periodic fluctuation characteristic, wherein the method comprises the following steps of:
presetting a sequence number weight parameter column matrix Q, a key value weight parameter column matrix K and a numerical weight parameter column matrix V;
and performing first self-attention encoding processing on the parameter distribution matrix B by using the sequence number weight parameter column matrix Q, the key value weight parameter column matrix K and the numerical value weight parameter column matrix V to obtain a periodic variation characteristic matrix C, wherein the calculation formula of the periodic variation characteristic matrix C is as follows:
C＝{C i }
wherein C is i An ith element representing the periodically varying feature matrix C; k (K) T Representing a transposed matrix of the key value weight parameter column matrix K; d, d k The method comprises the steps of presetting a key value weight parameter; e is a natural constant;
obtaining a periodic column coefficient matrix W according to the sequence number weight parameter column matrix Q, the key value weight parameter column matrix K and the numerical value weight parameter column matrix V o ；
Using the periodic column coefficient matrix W o And performing second self-attention coding processing on the periodic variation characteristic matrix C to obtain a periodic fluctuation quantity matrix D.
Further, based on the historical traffic data, the periodic variation characteristic and the periodic fluctuation characteristic, performing iterative training on a human demand prediction model to be trained to obtain a trained human demand prediction model, including:
inputting the input data matrix alpha, the periodic characteristic change matrix C and the periodic fluctuation quantity matrix D into a prediction model phi to be trained, and calculating to obtain a human demand training result x, wherein the calculation formula of the human demand training result x is as follows:
x＝Φ(α,C,D)＝θ[{max(0,wα+b)}+{C,D}]
wherein max is a preset full link layer activation function; w is a coefficient feature matrix; b is a preset induced deviation residual error; θ is a preset model parameter;
according to the human demand training result x, the coefficient feature matrix w is adjusted until the optimal human demand training result x is predicted by fitting the prediction model phi;
And obtaining the human demand prediction model after training by using the coefficient characteristic matrix w after adjustment.
Further, according to the historical traffic data, scheduling traffic data of a period to be scheduled is obtained through prediction, the scheduling traffic data is input into the trained human demand prediction model, a human demand prediction result is obtained, and according to the human demand prediction result, a first scheduling result is generated, and the method comprises the following steps:
according to historical synchronous traffic data and recent traffic data in the historical traffic data, predicting to obtain scheduling traffic data of a period to be scheduled, wherein the scheduling traffic data comprises at least one predicted incoming call request quantity, predicted call completing rate, predicted personnel attendance rate and predicted personnel utilization rate of a preset period;
inputting the scheduling telephone traffic data into the trained human demand prediction model, and outputting a corresponding human demand prediction result y;
matching the human demand prediction result y corresponding to each preset period with a preset working schedule to obtain a human demand list of each working time period of the period to be scheduled;
for each shift group, performing group allocation on staff according to the manual demand list of each working time period to obtain the number n of shift group workers of the shift group in each working time period j The number n of team workers in each working period j The calculation formula of (2) is as follows:
wherein k is the total preset cycle number of the period to be scheduled; h is a j The staff utilization rate is preset; y is j The human demand prediction results of the corresponding working time periods in the human demand list of each working time period are obtained; j is greater than or equal to 1 and less than or equal to kAn integer;
team working number n using each working time period j The manual demand list of each working time period is used for obtaining the shift and change results of each working time period;
generating a personal scheduling result of each staff according to the group rotation result of each working time period;
and taking the shift results of each working time period group and the personal shift result as the first shift result.
Further, performing verification analysis processing on the first shift result to obtain a verification analysis result, and if the verification analysis result is qualified, taking the first shift result as a second shift result, including:
aiming at the group staff number of each shift group, calculating the difference value between the group staff number of the current group and the group staff numbers of the rest groups to obtain a group staff number difference value;
judging whether the difference value of the number of the teams is larger than a preset number difference value;
If the difference value of the number of the teams is larger than the preset difference value of the number of the teams, the verification result of the number of the teams is that the teams do not pass;
if the difference value of the number of the teams is not larger than the preset difference value of the number of the teams, the verification result of the number of the teams is passed;
counting to obtain the total scheduling working time of all staff by using all the personal scheduling results;
obtaining the average working time of the scheduling according to the total scheduling working time and the total number of workers;
aiming at each staff, the corresponding personal scheduling result is utilized to obtain the scheduling personal working time;
subtracting the working time of the shift individual from the average working time of the shift, and processing the absolute value of the operation result to obtain the absolute value of the working time difference;
judging whether the absolute value of the working time difference value is smaller than or equal to a preset working time threshold value;
if the absolute value of the working time difference is smaller than or equal to the preset working time threshold, the staff working time length verification result is passed;
if the absolute value of the working time difference is larger than the preset working time threshold, the staff working time length verification result is that the staff working time length verification result is not passed;
judging whether the team number verification result and the employee working time verification result are both passed;
If the team number verification result and the staff work time verification result are both passed, the first scheduling result is used as the second scheduling result;
if the verification result of the number of the teams and/or the verification result of the working time of the staff is not passed, discarding the first scheduling result and performing scheduling again.
According to a third aspect of the embodiments of the present invention, there is provided an intelligent shift schedule prediction apparatus, the apparatus including: a processor and a memory;
the memory is used for storing one or more program instructions;
the processor is configured to execute one or more program instructions to perform the steps of an intelligent scheduling timing prediction method as set forth in any one of the preceding claims.
According to a fourth aspect of embodiments of the present invention, there is provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of an intelligent scheduling timing prediction method as defined in any one of the above.
The embodiment of the invention has the following advantages:
the embodiment of the invention discloses an intelligent scheduling time sequence prediction method, a system, equipment and a storage medium, wherein, firstly, historical telephone traffic data of a call center is obtained according to a preset period; then, based on time sequence distribution characteristics of historical telephone traffic data, carrying out characteristic analysis on the historical telephone traffic data to obtain corresponding data change characteristics; then, the data change characteristics are self-attentively encoded by utilizing the periodic change rule of the data change characteristics, so that the corresponding periodic characteristics are obtained; training to obtain a human demand prediction model based on historical telephone traffic data and corresponding periodic characteristics; and generating a first scheduling result by using a human demand prediction result predicted by the human demand prediction model, performing verification analysis on the first scheduling result, and taking the verified first scheduling result as a second scheduling result. The embodiment of the invention effectively improves the practicability of the scheduling result, realizes the timely update of the scheduling result, and simultaneously avoids the problem of overlarge working hour difference among staff.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those of ordinary skill in the art that the drawings in the following description are exemplary only and that other implementations can be obtained from the extensions of the drawings provided without inventive effort.
The structures, proportions, sizes, etc. shown in the present specification are shown only for the purposes of illustration and description, and are not intended to limit the scope of the invention, which is defined by the claims, so that any structural modifications, changes in proportions, or adjustments of sizes, which do not affect the efficacy or the achievement of the present invention, should fall within the ambit of the technical disclosure.
FIG. 1 is a schematic diagram of a logic structure of an intelligent scheduling timing prediction system according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of an intelligent scheduling timing prediction method according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a human demand prediction model training process according to an embodiment of the present invention;
FIG. 4 is a schematic flow chart of the application of the human demand prediction model and the generation of the shift result according to the embodiment of the invention;
fig. 5 is a schematic flow chart of a shift result verification provided by an embodiment of the present invention.
Detailed Description
Other advantages and advantages of the present invention will become apparent to those skilled in the art from the following detailed description, which, by way of illustration, is to be read in connection with certain specific embodiments, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, an embodiment of the present invention provides an intelligent scheduling timing prediction system, which specifically includes: the system comprises a telephone traffic data acquisition module 1, a first data processing module 2, a second data processing module 3, a prediction model training module 4, a shift result generation module 5 and a shift result verification module 6.
Further, the telephone traffic data acquisition module 1 is used for acquiring historical telephone traffic data of the call center according to a preset period, wherein the historical telephone traffic data comprises a historical incoming call request quantity, a historical call completing rate, a historical personnel attendance rate and a historical personnel utilization rate; the first data processing module 2 is used for performing first characteristic analysis processing on the historical telephone traffic data based on the time sequence distribution characteristics of the historical telephone traffic data to obtain first data change characteristics; according to the data change rule of the first data change feature, performing second feature analysis processing on the first data change feature to obtain a second data change feature; the second data processing module 3 is configured to perform a first self-attention encoding process on the second data change feature according to a periodic change rule of the second data change feature, so as to obtain a corresponding periodic change feature; performing second self-attention coding processing on the periodic variation characteristics to obtain periodic fluctuation characteristics; the prediction model training module 4 is used for carrying out iterative training on the human demand prediction model to be trained based on historical telephone traffic data, periodic variation characteristics and periodic fluctuation characteristics to obtain a trained human demand prediction model; the scheduling result generating module 5 is used for predicting and obtaining scheduling traffic data of a to-be-scheduled period according to the historical traffic data, inputting the scheduling traffic data into a trained human demand prediction model to obtain a human demand prediction result, and generating a first scheduling result according to the human demand prediction result; the shift result verification module 6 is configured to perform verification analysis processing on the first shift result, obtain a verification analysis result, and if the verification analysis result is qualified, take the first shift result as a second shift result.
The embodiment of the invention discloses an intelligent scheduling time sequence prediction system, which comprises the steps of firstly, acquiring historical telephone traffic data of a call center according to a preset period; then, based on time sequence distribution characteristics of historical telephone traffic data, carrying out characteristic analysis on the historical telephone traffic data to obtain corresponding data change characteristics; then, the data change characteristics are self-attentively encoded by utilizing the periodic change rule of the data change characteristics, so that the corresponding periodic characteristics are obtained; training to obtain a human demand prediction model based on historical telephone traffic data and corresponding periodic characteristics; and generating a first scheduling result by using a human demand prediction result predicted by the human demand prediction model, performing verification analysis on the first scheduling result, and taking the verified first scheduling result as a second scheduling result. The embodiment of the invention effectively improves the practicability of the scheduling result, realizes the timely update of the scheduling result, and simultaneously avoids the problem of overlarge working hour difference among staff.
Corresponding to the intelligent scheduling time sequence prediction system disclosed by the invention, the embodiment of the invention also discloses an intelligent scheduling time sequence prediction method. The following describes in detail an intelligent scheduling timing prediction method disclosed in the embodiment of the present invention in conjunction with an intelligent scheduling timing prediction system described above.
With reference to fig. 2, specific steps of an intelligent scheduling timing prediction method provided by an embodiment of the present invention are described below.
The telephone traffic data acquisition module 1 acquires historical telephone traffic data of the call center according to a preset period, wherein the historical telephone traffic data comprises historical call-in request quantity, historical call completing rate, historical personnel attendance rate and historical employee utilization rate.
The steps specifically comprise: and taking half an hour as a preset period, counting the historical incoming call request quantity, the historical call completing rate, the historical personnel attendance rate and the historical personnel utilization rate of the call center.
The embodiment of the invention carries out traffic data statistics based on a shorter period, can more accurately reflect the change rule of data, and can respond to burst data abnormality and special change matters in a short time in time.
The first data processing module 2 performs first characteristic analysis processing on the historical telephone traffic data based on the time sequence distribution characteristic of the historical telephone traffic data to obtain a first data change characteristic; and carrying out second characteristic analysis processing on the first data change characteristics according to the data change rule of the first data change characteristics to obtain second data change characteristics.
Referring to fig. 3, the steps specifically include: generating an input data matrix alpha according to the historical telephone traffic data by utilizing the historical incoming call request quantity, the historical call completing rate, the historical personnel attendance rate and the historical personnel utilization rate in each preset period, wherein each column of the input data matrix alpha consists of the historical incoming call request quantity, the historical call completing rate, the historical personnel attendance rate and the historical personnel utilization rate in one preset period; since the historical traffic data is counted by taking half an hour as a period, the input data matrix alpha has uniform time sequence distribution characteristics; performing probability density operation on the input data matrix alpha to obtain a probability density matrix A corresponding to the input data matrix alpha, wherein the calculation formula of the probability density matrix A is as follows:
wherein A is i An i-th element representing a probability density matrix a; alpha i The ith column of the input data matrix alpha is represented, namely, historical traffic data of a preset period;
is formed by alpha i Column matrix of arbitrary values within the value range, e.g. the value range of the above-mentioned historic personnel attendance is 0 to 100, then ∈ ->
Is any value between 0 and 100; />
The Euclidean distance between the maximum value and the average value in the input data matrix alpha is represented; Δα represents an increment of the input data matrix α; i is an integer greater than or equal to zero and less than the total number of historical traffic data periods.
And (3) carrying out integral operation on the probability density matrix A to obtain a parameter distribution matrix B, wherein the calculation formula of the parameter distribution matrix B is as follows:
B＝{B i }
wherein t is a preset period parameter.
The embodiment of the invention constructs an input data matrix by utilizing the historical call request quantity, the historical call completing rate, the historical personnel attendance rate and the historical personnel utilization rate of a plurality of preset periods, calculates to obtain a probability density matrix of the input data matrix, and then carries out integral operation on the probability density matrix to obtain a corresponding parameter distribution matrix; the parameter distribution matrix obtained through the steps can completely describe the data change rule of the historical telephone traffic data, and provides a data basis for subsequent self-attention code learning.
The second data processing module 3 performs a first self-attention encoding process on the second data change feature according to the periodic change rule of the second data change feature to obtain a corresponding periodic change feature; and performing second self-attention coding processing on the periodic variation characteristic to obtain a periodic fluctuation characteristic.
Referring to fig. 3, the steps specifically include: presetting a sequence number weight parameter column matrix Q, a key value weight parameter column matrix K and a numerical weight parameter column matrix V; and performing first self-attention coding processing on the parameter distribution matrix B by using the sequence number weight parameter column matrix Q, the key value weight parameter column matrix K and the number weight parameter column matrix V to obtain a periodic variation characteristic matrix C, wherein the calculation formula of the periodic variation characteristic matrix C is as follows:
C＝{C i }
Wherein C is i An ith element representing a periodically varying feature matrix C; k (K) T A transpose matrix representing a key value weight parameter column matrix K; d, d k The method comprises the steps of presetting a key value weight parameter; e is a natural constant;
obtaining a periodic column coefficient matrix W according to the sequence number weight parameter column matrix Q, the key value weight parameter column matrix K and the numerical weight parameter column matrix V o The method comprises the steps of carrying out a first treatment on the surface of the Using a periodic column coefficient matrix W o And performing second self-attention coding processing on the periodic variation characteristic matrix C to obtain a periodic fluctuation quantity matrix D, wherein the calculation formula of the periodic fluctuation quantity matrix D is as follows:
D＝MultiHeadAttention(Q,K,V)＝W o C
wherein multi-head attention represents a multi-head attention codec learning function.
The embodiment of the invention encodes a parameter distribution matrix corresponding to historical telephone traffic data based on a self-attention mechanism to obtain a periodic variation characteristic matrix and a periodic fluctuation matrix; the correlation between the historical telephone traffic data is enhanced, and the efficiency of the subsequent model training is improved.
And the prediction model training module 4 carries out iterative training on the human demand prediction model to be trained based on the historical telephone traffic data, the periodic variation characteristics and the periodic fluctuation characteristics to obtain a trained human demand prediction model.
Referring to fig. 3, the steps specifically include: the input data matrix alpha, the periodic characteristic change matrix C and the periodic fluctuation matrix D are used as training data to be input into a prediction model phi to be trained for model training, a human demand training result x is output, and a calculation formula of the human demand training result x is as follows:
x＝Φ(α,C,D)＝θ[{max(0,wα+b)}+{C,D}]
wherein max is a preset full link layer activation function; w is a coefficient feature matrix; b is a preset induced deviation residual error; θ is a preset model parameter;
according to a human demand training result x output by the model, adjusting a coefficient feature matrix w in the model until a prediction model phi is fitted to predict an optimal human demand training result x; and obtaining a human demand prediction model after training according to the coefficient characteristic matrix w after adjustment.
According to the embodiment of the invention, historical telephone traffic data is used as model training data, a periodic characteristic change matrix and a periodic fluctuation quantity matrix are used as training characteristics, a prediction model is trained, and the integral dissimilarity and contemporaneous local dissimilarity of the change with time are analyzed by utilizing the internal relation between the data, so that a trained manpower demand prediction model is obtained; the model can predict the number of manpower demands in a period of time according to traffic data in the period of time, and is used for subsequent shifts.
And the scheduling result generating module 5 predicts the scheduling traffic data of the period to be scheduled according to the historical traffic data, inputs the scheduling traffic data into the trained human demand prediction model to obtain a human demand prediction result, and generates a first scheduling result according to the human demand prediction result.
Referring to fig. 4, the steps specifically include: determining a to-be-scheduled time period for scheduling according to service requirements, and predicting scheduling traffic data of the to-be-scheduled time period based on historical traffic data of the same period in the past year by referring to recent traffic data; inputting the scheduling traffic data into a trained human demand prediction model, and outputting a human demand prediction result t of each preset period in a period to be scheduled; matching each human demand prediction result t with a preset working time table manufactured according to the period to be shifted to obtain a human demand table of each working time period of the period to be shifted; for each shift group, performing group allocation on staff according to the manual demand table of each working time period to obtain each shift group in each workerTeam working number n of working time period j Team working number n in each working period j The calculation formula of (2) is as follows:
wherein k is the total preset cycle number of the period to be scheduled; h is a j The staff utilization rate is preset; y is j The human demand prediction results of the corresponding working time periods in the human demand list of each working time period are obtained; j is an integer greater than or equal to 1 and less than or equal to k;
team working number n according to each working time period j And a manual demand table of each working time period to obtain a team rotation result delta of each working time period j The shift result delta of each working period shift group j The calculation formula of (2) is as follows:
shift result delta according to each working time period j And generating a personal scheduling result of each staff member, and taking the shift result of each working time period group and the personal scheduling result as a first scheduling result.
And the first scheduling result is subjected to verification analysis processing by the scheduling result verification module 6 to obtain a verification analysis result, and if the verification analysis result is qualified, the first scheduling result is used as a second scheduling result.
Referring to fig. 5, the steps specifically include: firstly, for the group staff number of each shift group, calculating the difference between the current group staff number of the current group and the group staff numbers of the rest groups to obtain a group staff number difference; judging whether the difference value of the number of groups of people is larger than the preset number of people; if the difference value of the number of the teams is larger than the preset difference value of the number of the teams, the verification result of the number of the teams is that the teams do not pass; if the difference value of the number of the teams is not larger than the preset difference value of the number of the teams, the verification result of the number of the teams is passed; counting to obtain the total scheduling working time of all staff by using all personal scheduling results; dividing the total scheduling working time by the total number of workers to obtain the average scheduling working time; aiming at each staff, the corresponding personal scheduling result is utilized to obtain the scheduling personal working time; then subtracting the working time of the person on the shift and the average working time of the shift, and processing the absolute value of the operation result to obtain the absolute value of the working time difference; judging whether the absolute value of the working time difference value is smaller than or equal to a preset working time threshold value; if the absolute value of the working time difference is smaller than or equal to the preset working time threshold, the staff working time length verification result is passed; if the absolute value of the working time difference is larger than the preset working time threshold, the staff working time length verification result is not passed; judging whether the team number verification result and the staff working time verification result are both passed; if the team number verification result and the staff working time verification result are both passed, the first scheduling result is used as a second scheduling result; if the verification result of the number of the teams and/or the verification result of the working time of the staff are not passed, discarding the first scheduling result and performing scheduling again; and publishing the second scheduling result.
The embodiment of the invention discloses an intelligent scheduling time sequence prediction method, which comprises the steps of firstly, acquiring historical telephone traffic data of a call center according to a preset period; then, based on time sequence distribution characteristics of historical telephone traffic data, carrying out characteristic analysis on the historical telephone traffic data to obtain corresponding data change characteristics; then, the data change characteristics are self-attentively encoded by utilizing the periodic change rule of the data change characteristics, so that the corresponding periodic characteristics are obtained; training to obtain a human demand prediction model based on historical telephone traffic data and corresponding periodic characteristics; and generating a first scheduling result by using a human demand prediction result predicted by the human demand prediction model, performing verification analysis on the first scheduling result, and taking the verified first scheduling result as a second scheduling result. The embodiment of the invention effectively improves the practicability of the scheduling result, realizes the timely update of the scheduling result, and simultaneously avoids the problem of overlarge working hour difference among staff.
In addition, the embodiment of the invention also provides intelligent scheduling time sequence prediction equipment, which comprises: a processor and a memory; the memory is used for storing one or more program instructions; the processor is configured to execute one or more program instructions to perform the steps of an intelligent scheduling timing prediction method as set forth in any one of the preceding claims.
In addition, the embodiment of the invention also provides a computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, and the computer program realizes the steps of the intelligent scheduling time sequence prediction method according to any one of the above steps when being executed by a processor.
In the embodiment of the invention, the processor may be an integrated circuit chip with signal processing capability. The processor may be a general purpose processor, a digital signal processor (Digital Signal Processor, DSP for short), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC for short), a field programmable gate array (Field Programmable GateArray, FPGA for short), or other programmable logic device, discrete gate or transistor logic device, discrete hardware components.
The disclosed methods, steps, and logic blocks in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be embodied directly in the execution of a hardware decoding processor, or in the execution of a combination of hardware and software modules in a decoding processor. The software modules may be located in a random access memory, flash memory, read only memory, programmable read only memory, or electrically erasable programmable memory, registers, etc. as well known in the art. The processor reads the information in the storage medium and, in combination with its hardware, performs the steps of the above method.
The storage medium may be memory, for example, may be volatile memory or nonvolatile memory, or may include both volatile and nonvolatile memory.
The nonvolatile Memory may be a Read-Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an electrically Erasable ROM (Electrically EPROM, EEPROM), or a flash Memory.
The volatile memory may be a random access memory (Random Access Memory, RAM for short) which acts as an external cache. By way of example, and not limitation, many forms of RAM are available, such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (Double Data Rate SDRAM), enhanced SDRAM (ESDRAM), synchronous DRAM (SLDRAM), and direct memory bus RAM (Direct Rambus RAM, DRRAM).
The storage media described in embodiments of the present invention are intended to comprise, without being limited to, these and any other suitable types of memory.
Those skilled in the art will appreciate that in one or more of the examples described above, the functions described in the present invention may be implemented in a combination of hardware and software. When the software is applied, the corresponding functions may be stored in a computer-readable medium or transmitted as one or more instructions or code on the computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media may be any available media that can be accessed by a general purpose or special purpose computer.
While the invention has been described in detail in the foregoing general description and specific examples, it will be apparent to those skilled in the art that modifications and improvements can be made thereto. Accordingly, such modifications or improvements may be made without departing from the spirit of the invention and are intended to be within the scope of the invention as claimed.
Claims (10)
1. An intelligent shift schedule prediction method, which is characterized by comprising the following steps:
According to a preset period, acquiring historical telephone traffic data of a call center, wherein the historical telephone traffic data comprises a historical call-in request quantity, a historical call completing rate, a historical personnel attendance rate and a historical personnel utilization rate;
based on the time sequence distribution characteristics of the historical telephone traffic data, performing first characteristic analysis processing on the historical telephone traffic data to obtain first data change characteristics; performing second characteristic analysis processing on the first data change characteristic according to the data change rule of the first data change characteristic to obtain a second data change characteristic;
according to the periodic variation rule of the second data variation characteristic, performing first self-attention coding processing on the second data variation characteristic to obtain a corresponding periodic variation characteristic; performing second self-attention coding processing on the periodic variation characteristics to obtain periodic fluctuation characteristics;
performing iterative training on the human demand prediction model to be trained based on the historical telephone traffic data, the periodic variation characteristics and the periodic fluctuation characteristics to obtain a trained human demand prediction model;
according to the historical telephone traffic data, scheduling telephone traffic data of a period to be scheduled is obtained through prediction, the scheduling telephone traffic data is input into the trained human demand prediction model, a human demand prediction result is obtained, and a first scheduling result is generated according to the human demand prediction result;
And carrying out verification analysis processing on the first scheduling result to obtain a verification analysis result, and taking the first scheduling result as a second scheduling result if the verification analysis result is qualified.
2. The intelligent shift schedule prediction method of claim 1, wherein performing a first feature analysis process on the historical traffic data based on the time sequence distribution feature of the historical traffic data to obtain a first data change feature comprises:
according to the historical telephone traffic data, an input data matrix alpha is obtained, and each column of the input data matrix alpha comprises the historical incoming call request quantity, the historical call completing rate, the historical personnel attendance rate and the historical personnel utilization rate in one preset period;
based on the time sequence distribution characteristics of the historical telephone traffic data, carrying out probability density operation on the input data matrix alpha to obtain a probability density matrix A corresponding to the input data matrix alpha, wherein the calculation formula of the probability density matrix A is as follows:
A＝{A i }
wherein A is i An i-th element representing the probability density matrix a; alpha i An ith column representing the input data matrix α;
is formed by alpha i A column matrix formed by any value in the value range; / >
Representing the Euclidean distance between the maximum value and the average value in the input data matrix alpha; Δα represents an increment of the input data matrix α; i is an integer greater than or equal to zero and less than the total number of historical traffic data periods.
3. The intelligent shift schedule prediction method according to claim 2, wherein the performing a second feature analysis on the first data change feature according to the data change rule of the first data change feature to obtain a second data change feature comprises:
performing integral operation on the probability density matrix A to obtain a parameter distribution matrix B, wherein the calculation formula of the parameter distribution matrix B is as follows:
B＝{B i }
wherein t is a preset period parameter.
4. The intelligent shift schedule prediction method according to claim 3, wherein the first self-attention encoding process is performed on the second data change feature according to the periodic change rule of the second data change feature to obtain a corresponding periodic change feature; and performing second self-attention coding processing on the periodic variation characteristic to obtain a periodic fluctuation characteristic, wherein the method comprises the following steps of:
presetting a sequence number weight parameter column matrix Q, a key value weight parameter column matrix K and a numerical weight parameter column matrix V;
And performing first self-attention encoding processing on the parameter distribution matrix B by using the sequence number weight parameter column matrix Q, the key value weight parameter column matrix K and the numerical value weight parameter column matrix V to obtain a periodic variation characteristic matrix C, wherein the calculation formula of the periodic variation characteristic matrix C is as follows:
C＝{C i }
wherein C is i An ith element representing the periodically varying feature matrix C; k (K) T Representing a transposed matrix of the key value weight parameter column matrix K; d, d k The method comprises the steps of presetting a key value weight parameter; e is a natural constant;
obtaining a periodic column coefficient matrix W according to the sequence number weight parameter column matrix Q, the key value weight parameter column matrix K and the numerical value weight parameter column matrix V o ；
Using the periodic column coefficient matrix W o For the periodic variationThe sign matrix C performs a second self-attention encoding process to obtain a periodic fluctuation matrix D.
5. The intelligent shift schedule prediction method of claim 4, wherein iteratively training a human demand prediction model to be trained based on the historical traffic data, the periodic variation characteristics and the periodic fluctuation characteristics to obtain a trained human demand prediction model, comprising:
Inputting the input data matrix alpha, the periodic characteristic change matrix C and the periodic fluctuation quantity matrix D into a prediction model phi to be trained, and calculating to obtain a human demand training result x, wherein the calculation formula of the human demand training result x is as follows:
x＝Φ(α,C,D)＝θ[{max(0,wα+b)}+{C,D}]
wherein max is a preset full link layer activation function; w is a coefficient feature matrix; b is a preset induced deviation residual error; θ is a preset model parameter;
according to the human demand training result x, the coefficient feature matrix w is adjusted until the optimal human demand training result x is predicted by fitting the prediction model phi;
and obtaining the human demand prediction model after training by using the coefficient characteristic matrix w after adjustment.
6. The intelligent scheduling timing prediction method according to claim 5, wherein the scheduling traffic data of the period to be scheduled is predicted according to the historical traffic data, the scheduling traffic data is input to the trained human demand prediction model to obtain a human demand prediction result, and the generating of the first scheduling result according to the human demand prediction result comprises:
according to historical synchronous traffic data and recent traffic data in the historical traffic data, predicting to obtain scheduling traffic data of a period to be scheduled, wherein the scheduling traffic data comprises at least one predicted incoming call request quantity, predicted call completing rate, predicted personnel attendance rate and predicted personnel utilization rate of a preset period;
Inputting the scheduling telephone traffic data into the trained human demand prediction model, and outputting a corresponding human demand prediction result y;
matching the human demand prediction result y corresponding to each preset period with a preset working schedule to obtain a human demand list of each working time period of the period to be scheduled;
for each shift group, performing group allocation on staff according to the manual demand list of each working time period to obtain the number n of shift group workers of the shift group in each working time period j The number n of team workers in each working period j The calculation formula of (2) is as follows:
wherein k is the total preset cycle number of the period to be scheduled; h is a j The staff utilization rate is preset; y is j The human demand prediction results of the corresponding working time periods in the human demand list of each working time period are obtained; j is an integer greater than or equal to 1 and less than or equal to k;
team working number n using each working time period j The manual demand list of each working time period is used for obtaining the shift and change results of each working time period;
generating a personal scheduling result of each staff according to the group rotation result of each working time period;
and taking the shift results of each working time period group and the personal shift result as the first shift result.
7. The intelligent shift schedule prediction method according to claim 6, wherein performing verification analysis processing on the first shift result to obtain a verification analysis result, and if the verification analysis result is qualified, taking the first shift result as a second shift result, including:
aiming at the group staff number of each shift group, calculating the difference value between the group staff number of the current group and the group staff numbers of the rest groups to obtain a group staff number difference value;
judging whether the difference value of the number of the teams is larger than a preset number difference value;
if the difference value of the number of the teams is larger than the preset difference value of the number of the teams, the verification result of the number of the teams is that the teams do not pass;
if the difference value of the number of the teams is not larger than the preset difference value of the number of the teams, the verification result of the number of the teams is passed;
counting to obtain the total scheduling working time of all staff by using all the personal scheduling results;
obtaining the average working time of the scheduling according to the total scheduling working time and the total number of workers;
aiming at each staff, the corresponding personal scheduling result is utilized to obtain the scheduling personal working time;
subtracting the working time of the shift individual from the average working time of the shift, and processing the absolute value of the operation result to obtain the absolute value of the working time difference;
Judging whether the absolute value of the working time difference value is smaller than or equal to a preset working time threshold value;
if the absolute value of the working time difference is smaller than or equal to the preset working time threshold, the staff working time length verification result is passed;
if the absolute value of the working time difference is larger than the preset working time threshold, the staff working time length verification result is that the staff working time length verification result is not passed;
judging whether the team number verification result and the employee working time verification result are both passed;
if the team number verification result and the staff work time verification result are both passed, the first scheduling result is used as the second scheduling result;
if the verification result of the number of the teams and/or the verification result of the working time of the staff is not passed, discarding the first scheduling result and performing scheduling again.
8. An intelligent shift schedule prediction system, the system comprising:
the telephone traffic data acquisition module is used for acquiring historical telephone traffic data of the call center according to a preset period, wherein the historical telephone traffic data comprises a historical incoming call request quantity, a historical call completing rate, a historical personnel attendance rate and a historical personnel utilization rate;
The first data processing module is used for carrying out first characteristic analysis processing on the historical telephone traffic data based on the time sequence distribution characteristics of the historical telephone traffic data to obtain first data change characteristics; performing second characteristic analysis processing on the first data change characteristic according to the data change rule of the first data change characteristic to obtain a second data change characteristic;
the second data processing module is used for carrying out first self-attention coding processing on the second data change characteristics according to the periodic change rule of the second data change characteristics to obtain corresponding periodic change characteristics; performing second self-attention coding processing on the periodic variation characteristics to obtain periodic fluctuation characteristics;
the prediction model training module is used for carrying out iterative training on the human demand prediction model to be trained based on the historical telephone traffic data, the periodic variation characteristics and the periodic fluctuation characteristics to obtain a trained human demand prediction model;
the scheduling result generation module is used for predicting and obtaining scheduling traffic data of a period to be scheduled according to the historical traffic data, inputting the scheduling traffic data into the trained human demand prediction model to obtain a human demand prediction result, and generating a first scheduling result according to the human demand prediction result;
And the shift result verification module is used for carrying out verification analysis processing on the first shift result to obtain a verification analysis result, and if the verification analysis result is qualified, the first shift result is used as a second shift result.
9. An intelligent shift schedule prediction apparatus, the apparatus comprising: a processor and a memory;
the memory is used for storing one or more program instructions;
the processor is configured to execute one or more program instructions for performing the steps of an intelligent scheduling timing prediction method according to any one of claims 1 to 7.
10. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the steps of an intelligent scheduling timing prediction method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211521463.3A CN116128558B (en) | 2022-11-30 | 2022-11-30 | Intelligent scheduling time sequence prediction method, system, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211521463.3A CN116128558B (en) | 2022-11-30 | 2022-11-30 | Intelligent scheduling time sequence prediction method, system, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116128558A true CN116128558A (en) | 2023-05-16 |
| CN116128558B CN116128558B (en) | 2024-02-13 |
Family
ID=86294666
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211521463.3A Active CN116128558B (en) | 2022-11-30 | 2022-11-30 | Intelligent scheduling time sequence prediction method, system, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116128558B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009150719A1 (en) * | 2008-06-10 | 2009-12-17 | 富士通株式会社 | Business support program, business support device, and business support method |
| CN107844915A (en) * | 2017-11-29 | 2018-03-27 | 信雅达系统工程股份有限公司 | A kind of automatic scheduling method of the call center based on traffic forecast |
| CN113159715A (en) * | 2021-04-06 | 2021-07-23 | 杭州远传新业科技有限公司 | Customer service seat scheduling method, system, electronic device and storage medium |
| CN113222238A (en) * | 2021-05-07 | 2021-08-06 | 哈尔滨工业大学 | Optimization method and system for shift arrangement of on-duty personnel of hub airport |
| CN114519610A (en) * | 2022-02-16 | 2022-05-20 | 支付宝(杭州)信息技术有限公司 | Information prediction method and device |
| US20220198372A1 (en) * | 2019-09-11 | 2022-06-23 | Hewlett-Packard Development Company, L.P. | Time-series machine learning model-based resource demand prediction |
| CN115049152A (en) * | 2022-07-12 | 2022-09-13 | 中国民用航空总局第二研究所 | Airport transfer area arrival passenger flow prediction method and device and electronic equipment |
-
2022
- 2022-11-30 CN CN202211521463.3A patent/CN116128558B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009150719A1 (en) * | 2008-06-10 | 2009-12-17 | 富士通株式会社 | Business support program, business support device, and business support method |
| CN107844915A (en) * | 2017-11-29 | 2018-03-27 | 信雅达系统工程股份有限公司 | A kind of automatic scheduling method of the call center based on traffic forecast |
| US20220198372A1 (en) * | 2019-09-11 | 2022-06-23 | Hewlett-Packard Development Company, L.P. | Time-series machine learning model-based resource demand prediction |
| CN113159715A (en) * | 2021-04-06 | 2021-07-23 | 杭州远传新业科技有限公司 | Customer service seat scheduling method, system, electronic device and storage medium |
| CN113222238A (en) * | 2021-05-07 | 2021-08-06 | 哈尔滨工业大学 | Optimization method and system for shift arrangement of on-duty personnel of hub airport |
| CN114519610A (en) * | 2022-02-16 | 2022-05-20 | 支付宝(杭州)信息技术有限公司 | Information prediction method and device |
| CN115049152A (en) * | 2022-07-12 | 2022-09-13 | 中国民用航空总局第二研究所 | Airport transfer area arrival passenger flow prediction method and device and electronic equipment |
Non-Patent Citations (2)
| Title |
|---|
| 唐迎春: ""G 公司呼叫服务排班管理优化策略研究"", 《中国优秀硕士学位论文全文数据库》 * |
| 孔梅娟: ""供电服务话务预测方法应用"", 《中国电力企业管理》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116128558B (en) | 2024-02-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shi et al. | A spatial–temporal attention approach for traffic prediction | |
| CN111080032B (en) | Load prediction method based on transducer structure | |
| Alencar et al. | Hybrid approach combining SARIMA and neural networks for multi-step ahead wind speed forecasting in Brazil | |
| Zhao et al. | Spatial-temporal aggregation graph convolution network for efficient mobile cellular traffic prediction | |
| CN111861023B (en) | Mixed wind power prediction method and device based on statistics | |
| WO2021129086A1 (en) | Traffic prediction method, device, and storage medium | |
| CN110009140B (en) | Daily power load prediction method and prediction device | |
| CN112508118B (en) | Target object behavior prediction method aiming at data offset and related equipment thereof | |
| EP4135264A1 (en) | Method and apparatus for predicting node state | |
| WO2021004324A1 (en) | Resource data processing method and apparatus, and computer device and storage medium | |
| Toubeau et al. | Capturing spatio-temporal dependencies in the probabilistic forecasting of distribution locational marginal prices | |
| He et al. | Probabilistic solar irradiance forecasting via a deep learning‐based hybrid approach | |
| WO2021000645A1 (en) | Contract generation method and apparatus based on data crawling and computer device | |
| CN114330935A (en) | New energy power prediction method and system based on multiple combined strategy integrated learning | |
| CN115600686A (en) | Personalized Transformer-based federal learning model training method and federal learning system | |
| CN116128558B (en) | Intelligent scheduling time sequence prediction method, system, equipment and storage medium | |
| CN113537591A (en) | Long-term weather prediction method and device, computer equipment and storage medium | |
| CN113223502A (en) | Speech recognition system optimization method, device, equipment and readable storage medium | |
| Nawar et al. | Transfer learning in deep learning models for building load forecasting: Case of limited data | |
| CN115730717A (en) | Power load prediction method and system based on combination of transfer learning strategy and multiple channels | |
| CN114925940A (en) | Holiday load prediction method and system based on load decomposition | |
| CN110753366A (en) | Prediction processing method and device for industry short message gateway capacity | |
| CN113505049A (en) | Cloud computing platform load prediction method based on firefly algorithm and BP neural network | |
| CN113449642A (en) | PM based on multi-scale feature extraction and deep learning2.5Concentration prediction method and system | |
| Khashman et al. | Oil price prediction using a supervised neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |