Browse Prior Art Database

Linear Programming Approach for the Allocation of GPRS/EDGE Data Services on to Radio Channels

IP.com Disclosure Number: IPCOM000010510D
Original Publication Date: 2003-Jan-25
Included in the Prior Art Database: 2003-Jan-25

Publishing Venue

Siemens

Related People

Juergen Carstens: CONTACT

Abstract

In mobile data communication different users compete for the same scare radio resources for voice and data services. Different data users request different target throughput rates (depending on the negotiated Quality of Service Profile, e.g. traffic class, service precedence, etc., and also depending on the contract of the subscriber with the operator, e.g. premium user, normal user, etc.). The degree a user is satisfied can be measured by the ratio of the data rate the user obtains divided by the data rate the user requested. Yet, the data rate a user obtains depends on how the available radio resources are subdivided among the competing users. Each user wants to have maximized its satisfaction level or utility. But since resources are scare, the maximation of the utility of one user decreases the utilities of all others. Thus there is the need for a “fairness principle” or a “joint utility function”, which allows to decide how the interests of the competing user can be balanced with each other. Individual (user) Utility function The utility function Ui of a user which is describing the level of user satisfaction is defined here as the ratio of the data rate the user will obtain due to the allocation of resources and the target throughput rate. Besides the throughput the utility function can be expanded towards other QoS (Quality of Service) principles such as delay requirements, error rates, etc., for simplicity we will limit the approach according to the user throughput in the following: The target throughput rate can be e.g. - requested mean throughput rate according to 3GPP R97/98 QoS parameters - guaranteed bit rate according to 3GPP R99 QoS parameters - a certain share of the maximum bit rate (according to 3GPP R99) or peak throughput rate (according to 3GPP R97/98) - a network defined target value equal for all services or dependent on the traffic class and/or traffic handling priority and/or allocation/retention priority (according to 3GPP R99) or dependent on the service precedence and/or radio priority (according to 3GPP R97/98) - a network defined target value according to subscription parameters Joint (network) Utility Functions To balance the competing requests of the individual users for the maximation of their utility functions Ui requires the construction of a “joint” or “all-user” utility function U = F(U1,...,UN) out of the individual utility functions Ui. The maximation of the “all-users” function F involves the maximation of the individual user’s levels together with the optimization of the inter-user distribution of the scare resources. The choice of F is restricted by rationality requirements. For each time t the resource allocation will be mathematically described by the “all-user” function U taking into account all running voice and packet data calls in a particular cell plus all new arrived calls. The optimum solution is achieved by solving a multidimensional optimization problem with hard and soft constraints.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 8% of the total text.

Page 1 of 11

S

© SIEMENS AG 2002 file: 2002J16473.doc page: 1

Linear Programming Approach for the Allocation of GPRS/EDGE Data Services on to Radio Channels

Idea: Franz Treml, DE-Munich; Dr. Carsten Ball, DE-Munich; Robert Muellner, DE-Munich

In mobile data communication different users compete for the same scare radio resources for voice and data services. Different data users request different target throughput rates (depending on the negotiated Quality of Service Profile, e.g. traffic class, service precedence, etc., and also depending on the contract of the subscriber with the operator, e.g. premium user, normal user, etc.). The degree a user is satisfied can be measured by the ratio of the data rate the user obtains divided by the data rate the user requested. Yet, the data rate a user obtains depends on how the available radio resources are subdivided among the competing users. Each user wants to have maximized its satisfaction level or utility. But since resources are scare, the maximation of the utility of one user decreases the utilities of all others. Thus there is the need for a "fairness principle" or a "joint utility function", which allows to decide how the interests of the competing user can be balanced with each other.

Individual (user) Utility function

The utility function Ui of a user <i> which is describing the level of user satisfaction is defined here as the ratio of the data rate the user will obtain due to the allocation of resources and the target throughput rate. Besides the throughput the utility function can be expanded towards other QoS (Quality of Service) principles such as delay requirements, error rates, etc., for simplicity we will limit the approach according to the user throughput in the following:

The target throughput rate can be e.g.

- requested mean throughput rate according to 3GPP R97/98 QoS parameters - guaranteed bit rate according to 3GPP R99 QoS parameters
- a certain share of the maximum bit rate (according to 3GPP R99) or peak throughput rate

  (according to 3GPP R97/98) - a network defined target value equal for all services or dependent on the traffic class and/or traffic handling priority and/or allocation/retention priority (according to 3GPP R99) or dependent on the service precedence and/or radio priority (according to 3GPP R97/98)

- a network defined target value according to subscription parameters

Joint (network) Utility Functions

To balance the competing requests of the individual users for the maximation of their utility functions Ui requires the construction of a "joint" or "all-user" utility function U = F(U1,...,UN) out of the individual utility functions Ui. The maximation of the "all-users" function F involves the maximation of the individual user's levels together with the optimization of the inter-user distribution of the scare resources. The choice of F is restricted by rationality requirements.

For each time t the resource allocation will be mathematically described by the "all-u...