Publication Date:
2020-08-05
Description:
The current work considers random access communications within a single cell and provides
an algorithmic suggestion for the almost optimal coordination of user actions.
In the scenario considered, an access effort is successful if (a) the signal is
detected at the receiver and (b) no collision occurs. The first event is controlled by the user transmission power and the second by the
choice of access (back-off) probability. These constitute the user action pair.
The algorithm aims at exploiting information from measurements and user reports, in order to estimate current
values of the system situation. Based on these, two optimization problems can be formulated and solved
for the so called contention level and transmission power level at the
base station side. The methodology to find the optimal values is based on minimization of a drift function.
The two values are broadcast in order to help the users update their actions ``almost optimally``.
In this way the wireless cell can achieve self-optimization, without outside coordination, by relying on such intelligent information exchange
and parameter estimation. Numerical results illustrate the great benefits of the
suggested algorithm, compared to scenarios where the actions remain fixed, at a very low or even zero cost in power expenditure and delay.
Language:
English
Type:
reportzib
,
doc-type:preprint
Format:
application/pdf
