System and method for calibrating a wfm scheduling module

Abstract

System and method for calibration of WFM system modeling parameters. A first mode M[D,S] of a modeler computes demand-shrinkage controlled service levels and an error metric e(M[D,S]) between the controlled and actual service levels. A user device iteratively adjusts each core parameter. When the user is satisfied that e(M[D,S]) is sufficiently small, calibration of the core parameters is complete. The same is done for calibrating the modeling factor, and then a final e(M[D,S])f is computed. A second mode M[D] computes, using the parameters just calibrated, demand-controlled service levels and an error metric e(M[D]) between the controlled service levels and actual levels. The user iteratively adjusts the shrinkage. When the user is satisfied that e(M[D]) is sufficiently small, calibration of the core parameters is complete.

1 . A method of calibrating a scheduling model parameters of a workforce management (WFM) scheduling model, the scheduling model parameters comprises user-set calibration parameters and a shrinkage, the method comprising:
 adjusting the user-set calibration parameters and the shrinkage, wherein the user-set calibration parameters comprise core parameters and modeling factors;   calculating for each adjustment of each of the user-set calibration parameters, a first service level error metric that is determined in accordance with actual demands and an actual schedule from a previous period and a shrinkage of zero to produce a calibrated user-set calibration parameters;   calculating for each adjustment of shrinkage, a second service level error metric that is determined in accordance with actual demands, and the calibrated user-set calibration parameters to produce a calibrated shrinkage;   updating a WFM database of a WFM system with a difference between a final first service level error metric and a third service level error metric, the calibrated user-set calibration parameters and the calibrated shrinkage.   
     
     
         2 . The method of  claim 1 , wherein the user-set calibration parameters are calibrated such that the core parameters are calibrated before the modeling factor. 
     
     
         3 . The method of  claim 1 , further comprising calculating a scheduling error metric e(M) that is determined in accordance with a predicted service levels and the actual service levels. 
     
     
         4 . The method of  claim 3 , further comprising displaying a calibration health indicator configured to indicate the scheduling error metric e(M). 
     
     
         5 . The method of  claim 1 , further comprising
 calculating a fourth calibration-error metric e(M[S]) that is determined in accordance with a forecasted demand, the actual schedule, the shrinkage set to zero, the calibrated user-set calibration parameters, and the actual service levels.   
     
     
         6 . The method of  claim 5 , wherein the fourth calibration-error metric is saved in a WFM database, an I/O device of a user of WFM system is alerted, or any combination thereof. 
     
     
         7 . The method of  claim 1 , further comprising:
 updating a calibration health indicator comprising:
 receiving, by a processor, predicted service levels and actual service levels from a workforce management system database; 
 calculating, by the processor, a scheduling error metric e(M) between the predicted service levels and the actual service levels; and 
 displaying, by the processor, on a calibration health indicator the difference between the scheduling error metric and a predetermined scheduling model error tolerance. 
   
     
     
         8 . The method of  claim 6 , wherein the calibration health indicator is displayed on a display in a WFM system. 
     
     
         9 . A method for calibrating scheduling model parameters of a work force management (WFM) scheduling model, comprising:
 obtaining, by a modeler in a schedule analyzer of a WFM system, modeling data for a previous period comprising:
 forecasted demands; 
 a published schedule; 
 predicted service levels; 
 actual demands; 
 an actual schedule; 
 actual service levels; and 
 scheduling model parameters comprising a shrinkage and user-set calibration parameters, wherein user-set calibration parameters comprise core parameters and a modeling factor; 
   executing the following calibration:
 instructing, by a calibration director and interface (CDI) of the WFM system, a statistics predictor and the modeler in a demand-shrinkage controlled mode, to compute, a first calibration-error metric e(M[D,S]) as a function of the actual demands, the actual schedule, the shrinkage set to zero, the core parameters, and the actual service levels, wherein the statistics predictor, and the modeler are comprised in the schedule analyzer; 
 receiving, by the CDI, the first calibration-error metric e(M[D,S]); and 
 displaying, by the CDI, the first calibration-error metric on an input/output device (I/O device) of the WFM system; 
   repeating the following process separately for each of user-set calibration parameters, until the CDI receives an input from the I/O device to stop adjusting for the user-set calibration parameters, wherein the input to stop is based on achieved calibration of the user-set calibration parameters to produce calibrated user-set calibration parameters:
 receiving, by the CDI from the I/O device, an adjustment to each of the user-set calibration parameters value; 
 transmitting, by the CDI to the modeler, the adjustment to each of the user-set calibration parameters value; 
 instructing, by the CDI, the modeler to compute a new first calibration-error metric e(M[D,S]), substituting adjustment value for those used in the computation of the first calibration-error metric; 
 receiving, by the CDI from the modeler, the new first calibration-error metric; and 
 displaying, by the CDI to the I/O device, the new first calibration-error metric; 
   instructing, by the CDI, the statistics predictor, and the modeler in a demand controlled mode, to compute, a second calibration-error metric as a function of the actual demand, the published schedule, the shrinkage, the calibrated core parameters, the calibrated modeling factor and the actual service levels;   receiving, by the CDI, the second calibration-error metric e(M[D]);   displaying, by the CDI, a calibration-error metric difference e(M[D,S])−e(M[D]) on the I/O device;   receiving, by the CDI from the I/O device, an adjustment to shrinkage;   transmitting, by the CDI, to the modeler the adjustment to shrinkage and the calibrated core parameter and the calibrated modeling factor;   repeating the following, until the CDI receives an input from the I/O device to stop adjusting, wherein the input to stop is based on calibration shrinkage is achieved to produce a calibrated shrinkage:
 receiving, by the CDI from the I/O device, an adjustment to the shrinkage; 
 transmitting, by the CDI to the modeler, the adjustment to the shrinkage; 
 instructing, by the CDI the modeler to compute a new second calibration-error metric e(M[D]), substituting the adjustment to the shrinkage for those used in the computation of the second calibration-error metric; 
 receiving, by the CDI from the modeler, the new second calibration-error metric; and 
 displaying, by the CDI to the I/O device, the new second calibration-error metric; and 
   updating a WFM database of the WFM system with the calibrated core parameters, the calibrated modeling factor, and the calibrated shrinkage, replacing the same from the previous period.   
     
     
         10 . The method of  claim 9 , wherein the user-set calibration parameters are calibrated such that the core parameters are calibrated before the modeling factor. 
     
     
         11 . The method of  claim 9 , further comprising:
 receiving, by the CDI, predicted service levels and the actual service levels;   calculating, by the CDI, a scheduling error metric e(M); and   displaying, by the CDI on the I/O device, a calibration health indicator configured to indicate the difference between the scheduling error metric and a predetermined scheduling model error tolerance.   
     
     
         12 . The method of  claim 9  wherein, the I/O device instructs the WFM system to start calibration of scheduling model parameters. 
     
     
         13 . The method of  claim 9 , wherein the CDI stores in the WFM database, the modeling data calibrated by the method as historical calibrated modeling data, and the first and second calibration-error metrics computed last that achieved calibration. 
     
     
         14 . The method of  claim 9 , further comprising
 instructing, by the CDI, a schedule optimizer, the statistics predictor and the modeler, in a shrinkage controlled mode, to compute, a third calibration-error metric e(M[S]) as a function of the forecasted demand, the actual schedule, a shrinkage set to zero, the calibrated core parameters, the calibrated modeling factor, and the actual service levels;   performing, by the CDI, one of the following: saving the third calibration-error metric in the WFM database, alerting an I/O device of a user of the WFM system, or any combination thereof.   
     
     
         15 . A system for calibrating modeling data of a work force management (WFM) system, comprising:
 a memory comprising instructions that when executed on a processor cause the system to perform a method comprising:   receive, by a modeler in a schedule analyzer of the WFM system, modeling data for a previous period comprising: forecasted demands; a published schedule; predicted service levels; actual demands; an actual schedule; actual service levels; and scheduling model parameters comprising a shrinkage and user-set calibration parameters, wherein user-set calibration parameters comprise core parameters and a modeling factor;   execute the following calibration:
 instruct, by a calibration director and interface (CDI) of the WFM system, a statistics predictor and the modeler in a demand-shrinkage controlled mode, to compute, a first calibration-error metric e(M[D,S]) as a function of the actual demands, the actual schedule, the shrinkage set to zero, the core parameters, and the actual service levels, wherein the statistics predictor, and the modeler are comprised in the schedule analyzer; 
 receive, by the CDI, the first calibration-error metric e(M[D,S]); 
 display, by the CDI, the first calibration-error metric on an input/output device (I/O device) of the WFM system; 
 repeat the following process separately for each of user-set calibration parameters, until the CDI receives an input from the I/O device to stop adjusting for the user-set calibration parameters, wherein the input to stop is based on achieved calibration of the user-set calibration parameters to produce calibrated user-set calibration parameters;
 (i) receive, by the CDI from the I/O device, an adjustment to each of the user-set calibration parameters value; 
 (ii) transmit, by the CDI to the modeler, the adjustment to each of the user-set calibration parameters value; 
 (iii) instruct, by the CDI, the modeler to compute a new first calibration-error metric e(M[D,S]), substituting adjustment value for those used in the computation of the first calibration-error metric; 
 (iv) receive, by the CDI from the modeler, the new first calibration-error metric; and 
 (v) display, by the CDI to the I/O device, the new first calibration-error metric; 
 
 instruct, by the CDI, the statistics predictor, and the modeler in a demand controlled mode, to compute, a second calibration-error metric as a function of the actual demand, the published schedule, the shrinkage, the calibrated core parameters, the calibrated modeling factor and the actual service levels; 
 receive, by the CDI, the second calibration-error metric e(M[D]); 
 display, by the CDI, a calibration-error metric difference e(M[D,S])−e(M[D]) on the I/O device; 
 receive, by the CDI from the I/O device, an adjustment to shrinkage; 
 transmit, by the CDI, to the modeler the adjustment to shrinkage and the calibrated core parameter and the calibrated modeling factor; 
 repeat the following, until the CDI receives an input from the I/O device to stop adjusting, wherein the input to stop is based on calibration shrinkage is achieved to produce a calibrated shrinkage;
 (i) receive, by the CDI from the I/O device, an adjustment to the shrinkage; 
 (ii) transmit, by the CDI to the modeler, the adjustment to the shrinkage; 
 (iii) instruct, by the CDI the modeler to compute a new second calibration-error metric e(M[D]), substituting the adjustment to the shrinkage for those used in the computation of the second calibration-error metric; 
 (iv) receive, by the CDI from the modeler, the new second calibration-error metric; and 
 (v) display, by the CDI to the I/O device, the new second calibration-error metric; and 
 
 update a WFM database of the WFM system with the calibrated core parameters, the calibrated modeling factor, and the calibrated shrinkage, replacing the same from the previous period. 
   
     
     
         16 . The system of  claim 15 , wherein the user-set calibration parameters are calibrated such that the core parameters are calibrated before the modeling factor. 
     
     
         17 . The system of  claim 15 , further comprising
 receive, by the CDI, predicted service levels and the actual service levels;   calculate, by the CDI, a scheduling error metric e(M); and   display, by the CDI on the I/O device, a calibration health indicator configured to indicate the difference between the scheduling error metric and a predetermined scheduling model error tolerance.   
     
     
         18 . The system of  claim 15 , wherein the CDI stores in the WFM database, the modeling data calibrated by the method as historical calibrated modeling data, and the first and second calibration-error metrics computed last that achieved calibration. 
     
     
         19 . The system of  claim 15 , further comprising
 instruct, by the CDI, a schedule optimizer, the statistics predictor and the modeler, in a shrinkage controlled mode, to compute, a third calibration-error metric e(M[S]) as a function of the forecasted demand, the actual schedule, a shrinkage set to zero, the calibrated core parameters, the calibrated modeling factor, and the actual service levels;   perform, by the CDI, one of the following: saving the third calibration-error metric in the WFM database, alerting an I/O device of a user of the WFM system, or any combination thereof.   
     
     
         20 . The system of  claim 15 , further comprising a calibration health indicator configured to display a scheduling model health level based on a calculated scheduling error metric, wherein the calculated scheduling error metric is determined in accordance with predicted service levels and the actual service levels.