Dynamic staffing of service centers to provide substantially zero-delay service

• US 6,330,326 B1
• Filed: 03/27/1998
• Issued: 12/11/2001
• Est. Priority Date: 03/27/1998
• Status: Expired due to Term

• Current Assignee: AT&T Corporation (AT&T, Inc.)
• Original Assignee: AT&T Corporation (AT&T, Inc.)
• Inventors: Whitt, Ward
• Primary Examiner(s): Hong, Harry S.
• Assistant Examiner(s): Deane, Jr., William J.
• Application Number:
  US09/049,037
• Time in Patent Office: 1,355 Days
• Field of Search: 379/265, 379/266, 379/309, 379/113, 379/219, 379/220, 379/265.13, 379/265.12, 379/266.02, 379/266.08
• US Class Current: 379/265.13
• CPC Class Codes: H04Q 2213/13072 Sequence circuits for call signaling, ACD systems
• CPC Class Codes: H04Q 2213/13093 Personal computer, PC
• CPC Class Codes: H04Q 2213/13103 Memory
• CPC Class Codes: H04Q 2213/13164 Traffic (registration, measurement,...)
• CPC Class Codes: H04Q 2213/13166 Fault prevention
• CPC Class Codes: H04Q 2213/13335 Simulation, emulation
• CPC Class Codes: H04Q 3/64 Distributing or queueing

Abstract

A service system provides substantially zero delay service and dynamically adjusts resources required to provide the service. According to an embodiment of the present invention, future staffing requirements of the service system are predicted by determining, of a number of customers currently in service, how many will remain in service at a predetermined future time and how many customers to arrive to the system in the future can be expected to remain in service at the predetermined future time. For customers in service, customers may be classified according to one or more attributes known for the customer. The attributes may be helpful to identify a type of service being provided to the customer and determine a remaining service time for the customer. Thus, the customer attributes may provide for more accurate staffing predictions than in the prior art.

Claims

1. A resource management method in a service system, comprising:

for each of a number of customers currently in service;

classifying the customer in service according to an attribute known for the customer, and

predicting, based on the attribute, a time when the customer in service will terminate service;

based on the predicted times of termination of the customers in service, estimating a number of the customers in service that will remain in service at a predetermined future time;

predicting a number of customers expected to arrive in the future that will remain in service at the predetermined future time; and

scheduling a number of agents sufficient to serve the predicted number of remaining customers in service and the predicted number of new customers at the predetermined future time.

2. The method of claim 1, wherein, for at least one customer in service, the time when the customer in service will terminate service is represented by a probability over time when the customer will terminate service.

3. The method of claim 1, wherein, for at least one customer in service, the time when the customer in service will terminate service is represented by a probability distribution function.

4. The method of claim 3, wherein the probability distribution function is based on an estimated mean time of service for the customer.

5. The method of claim 1, wherein, for at least one customer in service, the attribute of classification is a time elapsed since the one customer first began to receive service.

6. The method of claim 5, wherein, for the one customer, the time when the one customer will terminate service is represented by a conditional probability distribution function of a probability over time when the one customer will terminate service given the elapsed time for the one customer.

7. The method of claim 1, wherein, for at least one customer in service, the attribute of classification is the identity of an agent serving the one customer.

8. The method of claim 1, wherein, for at least one customer in service, the attribute of classification is an upper limit service time for the one customer.

9. The method of claim 1, wherein, for at least one customer in service, the attribute of classification is the identity of the one customer.

10. The method of claim 1, wherein, for at least one customer in service, the attribute of classification is conduct of the one customer during an elapsed time since the one customer first began to receive service.

11. A method of predicting resource requirement for a service system, comprising the steps of:

predicting a number of customers currently in service that will remain in service at a predetermined time,

predicting a number of future customers expected to arrive to the system before the predetermined time,

for each future customer, estimating a holding time of the future customer,

estimating a number of future customers that will remain in service at the predetermined time,

scheduling a number of agents sufficient to provide service to the number of current customers and the number of future customers that will remain in service at the predetermined time.

12. The method of claim 11, wherein, for at least one future customer, the time when the future customer will terminate service is represented by a probability over time when the customer will terminate service.

13. The method of claim 11, wherein, for at least one future customer, the time when the future customer will terminate service is represented by a probability distribution function.

14. The method of claim 13, wherein the probability distribution function is based on an estimated mean time of service for the future customer.

15. The method of claim 11, wherein the step of predicting a number of future customers includes predicting the number of future customers given an actual rate of arrival of customers to the service system.

16. The method of claim 11, wherein the step of predicting a number of future customers includes steps of:

predicting a number of future customers to arrive for each of a variety of call types and

estimating an aggregate number of future customers based on the predicted number of future customers of each call type.

17. The method of claim 11, wherein the step of predicting a number of future customers includes a prediction based on historical arrival rates of calls to the service system.

18. A resource management method, comprising:

for each source of current demand;

classifying the source according to an attribute known for the source, and

predicting, based on the attribute, a time when the source will cease its demand for service, based on the predicted times of cessation of the source of current demand, estimating a level of demand at a future time from the sources of demand expected to arrive in the future and remain at the future time, and

identifying an amount of service resources that will be required that will be required to meet the local level of demand in the future.