Section: New Results

Green Networking and Smart Grids

Participants : Sara Alouf, Eitan Altman, Alberto Benegiamo, Ioannis Dimitriou, Majed Haddad, Alain Jean-Marie, Giovanni Neglia.

Energy efficiency in wireless networks

In [25] , M. Haddad, P. Wiecek (Wroclaw Univ. of Technology, Poland), O. Habachi and Y. Hayel (both with Univ. of Avignon) investigated the achievable performances of multi-carrier energy efficient power control game. Both the simultaneous-move and the hierarchical games were addressed. For the first time, the analytical closed-form expressions of the spectrum coordination and the spectral efficiency of such models was derived. Results indicate that the spectrum coordination capability induced by the power control game model enables the wireless network to enjoy the energy efficiency improvement while still achieving a high spectral efficiency.

In [58] , the same authors studied energy efficiency of heterogeneous networks for both sparse and dense (two-tier and multi-tier) small cell deployments. The problem is formulated as a hierarchical (Stackelberg) game in which the macro cell is the leader whereas the small cell is the follower. Both players want to strategically decide on their power allocation policies in order to maximize the energy efficiency of their registered users. A backward induction method has been used to obtain a closed-form expression of the Stackelberg equilibrium. It was shown that the energy efficiency is maximized when only one sub-band is exploited for the players of the game depending on their fading channel gains.

In [34] , R. A. Vaca Ramirez and J. S. Thompson (Univ. of Edinburgh, UK), E. Altman and V. M. Ramos Ramos (Univ. Autonoma Metropolitana, Mexico) aim to reduce the power expenditure in the reverse link during low network load periods, by allocating extra resource blocks (RBs) to the mobile users. This is in contrast with other approaches in which resources are reduced in hours of low energy consumption. The user's rate demands are split among its allocated RBs in order to transmit in each of them by using a low level modulation order. In this low SINR regime the transmission is much more energy efficient since the log appearing in Shannon formula is in close to linear. We model the bandwidth expansion (BE) process by a game theory framework derived from the concept of stable marriage with incomplete lists (SMI).

P. Wiecek (Wroclaw Univ. of Technology, Poland) and E. Altman consider in [42] dynamic Multiple Access games between a random number of players competing over collision channels. Each of several mobiles involved in an interaction determines whether to transmit at a high or at a low power. High power decreases the lifetime of the battery but results in smaller collision probability. They formulated this game as an anonymous sequential game with undiscounted reward and computed the equilibrium [42] . The internal state of a player corresponds to the amount of energy left in the battery and the actions correspond to the transmission power.

I. Dimitriou investigated in [52] the power management of mobile devices, using a variant of an M/G/1 queue with probabilistic inhomogeneous multiple vacations and generalized service process. Under the vacation scheme, at the end of a vacation the server goes on another vacation, with a different probability distribution, if during the previous vacation there have been no arrivals. The modified vacation policy depends on the initial vacation interval and the server selects randomly over $M$ such vacation policies. The theoretical system can be applied for modeling the power saving mode of mobile devices in modern wireless systems. Moreover, the form of the service process properly describes the incremental redundancy retransmission scheme that provides different types of retransmissions in such systems. Steady state analysis is investigated, energy and performance metrics are obtained and used to provide numerical results that are also validated against simulations.

Energy efficiency in delay tolerant networks

Energy efficiency in mobile networks is further studied in [28] where L. Sassatelli (Univ. of Nice Sophia Antipolis), A. Ali, M. Panda and T. Chahed (all with Telecom SudParis) and E. Altman tackle the issue of reliable transport in Delay-Tolerant mobile ad hoc Networks, that are operated by some opportunistic routing algorithm. We propose a reliable transport mechanism that relies on Acknowledgements (ACK) and coding at the source. The various versions of the problem depending on buffer management policies are formulated, and a fluid model based on a mean-field approximation is derived for the designed reliable transport mechanism. This model allows to express both the mean file completion time and the energy consumption up to the delivery of the last ACK at the source.

Modeling of a smart green base station

S. Alouf, I. Dimitriou A. Jean-Marie have considered the modeling of wireless communication base stations with autonomous energy supply (solar, wind). They proposed and analyzed a queueing model to assess performance of a base station fully powered by renewable energy sources. The system operates in a finite state Markovian random environment that properly describes the intermittent nature of renewable energy sources and the data traffic. The base station is considered to be “smart” in the sense that it is able to dynamically adjust its coverage area, controlling thereby the traffic rate and its energy consumption. They show how the matrix-analytic formalism enables to construct and study the performance of a smart green base station operating in random environment. More precisely, the behavior of such a system is described by a five-dimensional Markov process, which is a homogeneous finite Quasi Birth-Death (QBD) process. Several existing algorithms can be used in order to obtain the stationary probability vector, which is the basis for the calculation of interesting performance metrics. This work is on-going and has not been submitted for publication yet.

Direct Load Control

Balancing energy demand and production is becoming a more and more challenging task for energy utilities also because of the larger penetration of renewable energies which are more difficult to predict and control. While the traditional solution is to dynamically adapt energy production to follow the time-varying demand, a new trend is to drive the demand itself. Most of the ongoing actions in this direction involve greedy energy consumers, like industrial plant, supermarkets or large buildings. Pervasive communication technologies may allow in the near future to push further the granularity of such approach, by having the energy utility interacting with residential appliances. In [65] and in its extension [64] , G. Neglia, in collaboration with G. Di Bella, L. Giarré and I. Tinnirello (Univ. of Palermo, Italy) study large scale direct control of inelastic home appliances whose energy demand cannot be shaped, but simply deferred. Their solution does not suppose any particular intelligence at the appliances. The actuators are rather smart plugs (simple devices with local communication capabilities that can be inserted between appliances plugs and power sockets) and are able to interrupt/reactivate power flow through the plug. A simple control message can be broadcast to a large set of smart plugs for probabilistically enabling or deferring the activation requests of a specific load type in order to satisfy a probabilistic bound on the aggregated power consumption. The control law and the most important performance metrics can be easily derived analytically.

Charge of Electric Vehicles

The massive introduction of Electric Vehicles (EVs) is expected to significantly increase the power load experienced by the electrical grid, but also to foster the exploitation of renewable energy sources: if the charge process of a fleet of EVs is scheduled by an intelligent entity such as a load aggregator, the EVs' batteries can contribute in flattening energy production peaks due to the intermittent production patterns of renewables by being recharged when energy production surpluses occur. To this aim, time varying energy prices are used, which can be diminished in case of excessive energy production to incentivize energy consumption (or increased in case of shortage to discourage energy utilization). In [70] G. Neglia, in cooperation with C. Rottondi and G. Verticale (Politecnico di Milano, Italy), evaluate the complexity of the optimal scheduling problem for a fleet of EVs aimed at minimizing the overall cost of the battery recharge in presence of time- variable energy tariffs. The scenario under consideration is a fleet owner having full knowledge of customers' traveling needs at the beginning of the scheduling horizon. They prove that the problem has polynomial complexity, provide complexity lower and upper bounds, and compare its performance to a benchmark approach which does not rely on prior knowledge of customers' requests, in order to evaluate whether the additional complexity required by the optimal scheduling strategy w.r.t. the benchmark is worthy the achieved economic advantages. Numerical results show considerable cost savings obtained by the optimal scheduling strategy.