Abstract

One important task in current and future communication networks is to define a suitable pricing scheme. It is then preferable to formulate a mathematical model, so that parameters will be optimized and important properties such as fairness or truthful anticipated load revelation (or incentive compatibility) will be verified. In this paper we study a simple and promising scheme called the cumulus pricing scheme, which can address service differentiation and scalability among other issues. Based on a mathematical model, we determine values for optimizing the provider’s revenue, which happens under the constraint that each user has an incentive to reveal its anticipated load. This has led to a small variation of the initial model from the literature as in the modelling, cumulus points are translated into financial terms, and measurements induce a cost as well.

Keywords: Cumulus pricing scheme; Network management; Quality of service

Article Outline

1. Introduction

2. The cumulus pricing scheme (CPS)

2.1. Model presentation

2.2. Mathematical definitions of revenues

2.3. CPS and incentives to cheat

3. The total penalty CPS

3.1. Definitions

3.2. About incentives to cheat in the total penalty CPS

4. Optimization of the provider’s revenue

4.1. Optimization problem and minimization of the cost function F_net

4.2. Optimization in a general setting

4.3. Analytic results in particular cases

4.3.1. Symmetric thresholds

4.3.2. Linear thresholds

4.3.3. Uniform thresholds

5. Numerical illustrations

5.1. Truthful anticipated consumption revelation

5.2. Simulated annealing for the general case

5.3. Special cases

5.3.1. Symmetric thresholds

5.3.2. Linear thresholds

5.3.3. Uniform threshold

6. Conclusions

References

Vitae