Economic theories can be used to efficiently use the resources of frequency, time, etc... in wireless networks while satisfying the needs of users described by Utility functions. In economics, Utility is a measure of satisfaction resulting from the consumption of goods. For example best effort traffic can be described by a logarithmic utility function in which users' satisfaction of the data rates decreases logarithmically with the increase of rates. This will be more efficient than simply asking for more rates which means a continuous demand of more resources with significantly adding to the satisfaction of users and may result in unfair and less efficient sharing of the resources among the different users.
The objective of this question is to get contributions explaining the use of Utility and other economic parameters in the optimization of wireless networks especially in systems using OFDM.
This certainly makes sense! In fact, the resource allocation in wireless networks is a multi-objective optimization problem, where the utility maximization of each user corresponds to the different objectives that one would like to maximize simulateously. Economic terminology such as Pareto optimality and utilitarism vs. egalitarianism have counterparts in radio resource allocation. This is something that we describe in our book "Optimal Resource Allocation in Coordinated Multi-Cell Systems" which you can download from here: http://kth.diva-portal.org/smash/get/diva2:608533/FULLTEXT01
You can also check out my presentation: http://www.ee.kth.se/~emilbjo/presentation_fnt-book.pptx
When it comes to different utility functions, the easiest thing would be to model it as some increasing function of the SINR. It can be logarithmic, convex, concave, S-shaped, etc. I give some different references to these in Section 1 of the book. In those cases, there is a lot of theory that can be used to solve the corresponding resource alloaction problem.
In general, it might be more approriate to talk about the quality-of-experience (QoE) which is not only a function of the SINR but also how much time it takes for a task (e.g., a transfer block of data) to finish and perhaps other properties. As far as I know, the modeling of QoE is an ongoing research topic with no clear consensus.
This certainly makes sense! In fact, the resource allocation in wireless networks is a multi-objective optimization problem, where the utility maximization of each user corresponds to the different objectives that one would like to maximize simulateously. Economic terminology such as Pareto optimality and utilitarism vs. egalitarianism have counterparts in radio resource allocation. This is something that we describe in our book "Optimal Resource Allocation in Coordinated Multi-Cell Systems" which you can download from here: http://kth.diva-portal.org/smash/get/diva2:608533/FULLTEXT01
You can also check out my presentation: http://www.ee.kth.se/~emilbjo/presentation_fnt-book.pptx
When it comes to different utility functions, the easiest thing would be to model it as some increasing function of the SINR. It can be logarithmic, convex, concave, S-shaped, etc. I give some different references to these in Section 1 of the book. In those cases, there is a lot of theory that can be used to solve the corresponding resource alloaction problem.
In general, it might be more approriate to talk about the quality-of-experience (QoE) which is not only a function of the SINR but also how much time it takes for a task (e.g., a transfer block of data) to finish and perhaps other properties. As far as I know, the modeling of QoE is an ongoing research topic with no clear consensus.
- Badges
- Science topic