This article has been accepted for inclusion in a future issue of this journal.

Content is final as presented, with the exception of pagination.


Supporting Mobile Cloud Computing in Smart Cities
via Randomized Algorithms
Daniela Mazza, Adela Pagès-Bernaus, Daniele Tarchi, Senior Member, IEEE, Angel Alejandro Juan,
and Giovanni Emanuele Corazza, Senior Member, IEEE

Abstract—Smart cities represent rich and dynamic environ- on the environment, cannot be neglected in a sustainable city
ments in which a multitude of smart mobile devices (SMDs) in- management [5].
teract among them by sharing data. SMDs require from fast access The increasing trends of both urbanization and mobile-
to online services, but they offer limited computing capabilities
and battery lifetime. SMDs make frequent use of computation of- connected people have driven the society toward the definition of
floading, delegating computing-intensive tasks to the cloud instead a geographic information system for smart cities that is charac-
of performing them locally. In such a large-scale and dynamic terized by the presence of a multitude of smart devices, sensors,
environment, there might be thousands of SMDs simultaneously and processing nodes aiming to improve urban services.
executing processes and, therefore, competing for the allotment of Smart mobile devices (SMDs) are becoming an essential part
remote resources. This arises the need for a smart allocation of these
resources. Accordingly, this paper proposes a biased-randomized of social life and the most effective and convenient communi-
algorithm to support efficient and fast link selection. This algo- cation tool, which is not bounded in time nor place. In recent
rithm is able to provide “real-time” near-optimal solutions that years, the increasing request of ubiquitous services has also led
outperform solutions obtained through existing greedy heuristics. to a sharp increase in the so-called cloud computing, where re-
Furthermore, it overcomes the responsiveness limitations of exact mote servers can be exploited for remote storage and processing
optimization methods.
services. This is even more important in smart city scenarios:
Index Terms—Biased-randomized algorithms, computation of- due to their complexity related to the number of users, hetero-
floading, HetNets, mobile cloud computing (MCC), smart city. geneous services, and specific requirements, smart cities can
significantly benefit from a cloud computing infrastructure as
I. INTRODUCTION it can be used to improve system performance and the battery
life of user devices through decreased work loads. Indeed, it
CCORDING to the flagship publication of the United Na-
A tion’s World Urbanization Prospect [1], more than one
half of the world population is currently living in urban areas,
is through information and communication technologies that
smart cities are truly turning smart [6], in particular by exploit-
ing SMDs, which together with cloud computing constitute the
and it is expected that about 70% of global population will be mobile cloud computing (MCC) framework [7]. A third element
city inhabitants by 2050. At the same time, the Visual Net- is the wireless network, allowing to link resource-constrained
working Index [2] outlines that mobile-connected devices have devices to centralized large data centers located inside the cloud.
already outnumbered the amount of people in the world. Includ- As suggested by Batty, to understand cities, we must view
ing machine-to-machine modules, there will be over 10 billion them not simply as places in space but as systems of networks
of such devices by 2018. Furthermore, an 11-fold increase in the and flows [8]. Thus, we can consider the MCC as a framework
overall mobile data traffic over the next five years is expected. that is the technological nervous system allowing the networks
Innovative smart city projects have been adopted in the po- and flows of the city to achieve a better urban way of life.
litical agenda of many governments as a key program to en- Moreover, the pervasiveness of wireless technologies has led to
able a vision where municipalities can use technology to meet the presence of heterogeneous networks where multiple types
sustainability goals, boost local economies, and improve ur- of access nodes operate simultaneously in the same city area
ban services. These projects focus on strengthening network [9], [10].
infrastructures, optimizing traffic, and transportation flows, de- In this context, one of the main challenges is to provide solu-
creasing energy consumption, and supplying innovative ser- tions able to jointly optimize the activities of data transfer (by
vices [3], [4]. In addition, a green concept of technology, in- exploiting wireless heterogeneous networks or HetNets) and
tended to mitigate and reverse the effects of human activity data processing (by delegating computing-intensive tasks to the
cloud in the MCC framework) [11], [12]. This strategy, com-
Manuscript received February 19, 2016; revised May 10, 2016; accepted May
28, 2016. This work was supported in part by the Spanish Ministry of Economy monly called cyberforaging or computation offloading, allows
and Competitiveness under Grant TRA2013-48180-C3-P and FEDER. us to overcome SMD limitations concerning limited battery
D. Mazza, D. Tarchi, and G. E. Corazza are with the Department of Electri- power and computational capacities. Moreover, it plays a key
cal, Electronic, and Information Engineering, University of Bologna, 40126
Bologna, Italy (e-mail:;; role in a smart environment where wireless communication is of utmost importance, particularly in mobility and traffic control
A. Pagès-Bernaus and A. A. Juan are with the Department of Computer domains [13].
Science–IN3, Open University of Catalonia, 08018 Barcelona, Spain (e-mail:; Although computation offloading can significantly increase
Digital Object Identifier 10.1109/JSYST.2016.2578358 data-processing capabilities of mobile users, it is challenging

1937-9234 © 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See standards/publications/rights/index.html for more information.

The remainder of this paper is organized as follows. ilar solutions to those provided by the optimal method. our proposed approach allows us to In particular. If a large number of Section VI. In [14]. and in [17]. packing. lems such as sequencing and scheduling problems [25]. quadratic assignment problems [27]. we propose a model establishing the application (viewed as an entity made out of data and operation) relationship among all the devices (the cloud infrastructure. first analyze the problem-related works in Section II and de. We biology and related fields. In fact. none of the above explicitly considers the interac- performance of the three approaches. . in among different devices and the cloud. present a dynamic offloading algorithm based on Lyapunov op- quirements and characteristics of the considered scenario. ination in the emerging mobile social networks. algorithm that allows us to find a near-optimal solution in short and partitions problems [29]. there are several works focused on the of- require excessive computation times. From this model. these methods cannot be used in practice. environment (macro and small cells). we show tive search procedure (GRASP)-like algorithms have been that it outclasses a greedy heuristic solution. computing times.g. Therefore. among others: schedul- mark. graph theory. investigate an energy-efficient offloading policy for transcod- ased or oriented random effect on the possible solutions of a ing as a service in a generic mobile cloud system.. a game-theoretic approach is proposed. RELATED WORKS leading to a reduced throughput for each user. representing an inefficient use of offloading operations [14]. location. clustering. this paper will focus on energy consumption and executed given an available wireless network connectivity. although aiming at reducing overall “community” or social costs. we present a biased-randomized cation and layout problems [28]. presented by the same authors in [16]. the greedy application models aiming to extract offloading friendly parts of heuristic. and very large scale integration design. the same All in all. as described in [30]. With respect to biased randomization techniques. we com. Finally. in this paper. In addition. and SMDs). The goal is tic algorithms—similar to the one presented here—have been to minimize the overall energy consumption by selecting the best widely used to solve many combinatorial optimization prob- communication link for each SMD. is performed by many users. In requesting devices. system consists of a dispatcher at the front end. These characteristics make focused on a global perspective. floading operation from the point of view of a single mobile user. Content is final as presented. They propose We will compare the solution provided by our probabilistic three algorithms for selecting the target set with only k users. the biased randomization techniques are described. lo- apply it in real time. greedy randomized adap- reach near-optimal solutions in “real time. Hereby. and summarized ing. the proposed random. showing that In this paper. In [10]. in [19] and [20]. Thus. Our approach in this paper is complexity of the greedy heuristic. the authors of a proper objective function taking into account different re. tion among users. In timization for selecting the software components to be remotely particular. with the exception of pagination. and the exact method—are based on the definition codes from existing applications [21]–[23]. This can lead to The cyberforaging mechanism for achieving efficient compu- a crucial increase in energy usage and data transmission times. since they In the literature. given that we have to routing problems [26]. corresponding to a centralized it an efficient alternative for real-life applications. the authors The biased-randomization techniques work by introducing a bi. transportation. given rized in Section IV-B was used for minimizing the collective cost that the modifications added do not increase the computational for a community in mobile clouds. trast. and a set of sible alternatives that are close to the global optimal. In [18]. In this problem setting. In con- in much less computing time. signment. routing. In [24]. the most promising node allow. covering. automatic drawing. manufacturing. time delays as key performance indicators for comparing the However. the focus of several works con- a selfish behavior in allocating resources to users. Net nodes. partitioning. power systems. Hence. probabilis- nication. Furthermore. 2 IEEE SYSTEMS JOURNAL to achieve an efficient coordination among the entire set of scribe the problem and the optimization model in Section III. telecommunications. vision that is more effective for a community. II. pare our approach to a greedy algorithm [16] that implements e. possible. This article has been accepted for inclusion in a future issue of this journal. citizens use the wireless resource to delegate computation to the cloud. the computation offloading entails data transmission Then. Such three cerning computation offloading is related to the development of techniques—the proposed probabilistic algorithm. tation offloading for MCC has been discussed in the literature. Moreover. We show that this cost func. Het. both in terms of computing and commu. the same authors proposed a utility function model ized algorithm clearly outperforms the greedy heuristic while derived from economics. a cost function is derived. The cloud problem. the efficiency of the access node can be severely affected: its channel capacity has to be shared among all the devices. in which the greedy heuristic summa- employing similar computing times once it is parallelized. applied to solve a wide set of problems. aware energy efficiency of the Base Stations composing the ing the potential increase in system efficiency has to be selected. in [16]. logic. This paper addresses this problem by con- The proposed randomized algorithm allows us to achieve sim.” Moreover. the authors biased-randomization techniques [15] to solve the link selection propose a scheme for optimizing the quality of service (QoS)- problem. for minimizing the mobile data traffic over cellular networks. as- in this paper for completeness. and in order to make the operation Section IV. the authors exploit set of alternative solutions is generated in short computational opportunistic communications to facilitate information dissem- times. allowing to choose the best solution from a set of pos. sidering entire populations instead of each user individually. a service engines at the back end. However. in Section V. vehicle tion cannot be optimally solved promptly. we present a probabilistic algorithm based on the game always admits a Nash equilibrium. the main findings and conclusions are highlighted. used as a bench. method with the optimal solution provided by exact methods. we present the numerical results.

By deploying additional small cells within the local-area range and bringing the network closer to the users. characterized by different features. The cloud entity Ccc is characterized by its own computa- The SMDs aim at delegating storage and computing tasks tional speed. the optimization problem is introduced. It combines a biased other application-focused works. such as RAT nodes. the authors propose the Iterated Local Search Efficient. . sumed by the cloud when sending the resultant data to the SMDs tribution is used to randomize the constructive process while [34]. Enscore. cloud computing infrastructure with a huge amount of storage in [32]. picocells. This article has been accepted for inclusion in a future issue of this journal. ering the computing processes to remote clusters. In the following. Content is final as presented. The ILS-ESP uses an iter. Ccc is a function of fcc . The cloud infrastructure Ccc refers to the presence of a remote tion process also guided by a geometric distribution. the entities involved in this scenario are keeping the logic behind the heuristic. mobile B. microcells. for accessing open data delivered by the public the permutation flow-Shop problem. Multiple RATs. the application can be computed lo- Fig. it is now possible to distribute the whole computational load. or skewed. leading to a and Ham heuristic [33] with a biased randomization process higher computing power. cally by the requesting SMD if this operation is not convenient. but it could suffer from III. the introduced. in amusement parks or traffic intersections). due to the data transfer from the users In this section. administration. considering the availability of direct Regarding the use of biased. HetNet nodes WiMAX. overall network capacities can be significantly boosted through a better reuse of the spatial resources. to a remote unit. and/or for storing big amounts of data. Different strategies can be The wireless HetNet infrastructure is composed of several carried out in order to increase the network capacity. the interaction in a pervasive environment (note that multihop links SimoRouting Generalized Clarke and Wrights Savings with are not taken into account in this paper). differently from the capacitated vehicle routing problem. Inspired by the attractive features and the potential advantages of the HetNets. or Ccc = Ccc (fcc ).: SUPPORTING MOBILE CLOUD COMPUTING IN SMART CITIES VIA RANDOMIZED ALGORITHMS 3 deployment of relays. Such network deployment is typically composed of a mix of low-power nodes underlying the conventional homogeneous macrocell network. 1. Hence. which include the energy con- randomization process with a base heuristic. HSPA+. It allows the citizens to interact Simple. randomization. and femtocells). sidered infinite. The proposed model focuses on direct connections between a RAT and multiple SMDs. such as IEEE 802. Beyond data storage. distributed antennas. space and computing power. and Parameter-Free (ILS-ESP) algorithm for solving remotely. Cloud heuristic for the arc routing problem with a biased randomiza. The centralized cloud allows to reduce the computational time by exploiting powerful processing units. Reference scenario with multiple SMDs in a multi-RAT environment. Furthermore.g. it does not represent a constraint in the putational times. The cloud infrastructure is often used for deliv- ated local search framework and combines the Nawaz. radio access technologies (RATs). due to the presence of a where several SMDs require cloud services.. This algorithm combines a savings-based A. it is possible to write that the common activities transferred to remote cloud infrastructures. fcc . Each SMD can single performing infrastructure leading to the single-point-of- connect to the cloud after selecting one of the available HetNet failure issue. Similarly. The storage availability of Ccc can be con- to overcome their resource constraints and to decrease com. However. e. as shown in Fig. due to the multiple focus is on an urban area with a pervasive wireless coverage. and LTE. users exploitation. PROBLEM DESCRIPTION AND OPTIMIZATION MODEL the distribution latency.g. are being integrated to create a heterogeneous wireless network. The to the cloud and vice versa. in [31].. the congestion. (1) just a fraction. authors developed the RandSHARP algorithm for solving the arc routing problem.g..11 WLANs. Likewise. The distributed smart city scenario is also an environment where the SMDs can choose and exploit different available RATs for transmitting the application data and for offloading toward the cloud..g. and the resiliency. MAZZA et al. thus. and small cellular base stations (e. their development has gained much momentum in the wireless industry and research communities in recent years. that is one of the most interaction with the SMDs. in residential homes and offices) and outdoors (e. with the exception of pagination. guided by a descending triangular distribution. 1. we con- Cache and Splitting algorithm is proposed in [15] for solving sider energy savings from the user-perspective. These deployments have occurred both indoors (e. A geometric dis.

l .. Considering an overall number of N devices and a total num- RAT = RAT(BW. Pid . . Content is final as presented. otherwise jth SMD. 4 IEEE SYSTEMS JOURNAL (1) BW : the nominal bandwidth of a certain communication Among different parameters that can affect the system perfor- technology that is available for the SMDs. and yj . cy )). j } + n i log2 {1 + SNRj e fcc (5) SNR: the signal-to-noise ratio at the receiver side of the   link between the SMD and the selected RAT node. we can resort to the Shannon formula for evaluating and ni is the number of SMDs connected to RATi . M . important when considering a scenario with mobile devices. the current and future needs of other users. SMDs application using the nodes of set M = {0. cy )). . . SNR. it is possible to (2) Pl : the power consumed by the SMD for local computa- write the energy consumed by SMDj when offloading through tion. . Ptr j D Pid j O (4) fm d : the speed to perform the computation locally. as (3) Pid : the power consumed by the SMD in waiting mode   while the computation is performed in the cloud. O. . aside from the application partitioning. j. its con- application that is characterized by the number of operations sumed energy is to be executed.j (cx . the RATi in (7). . defined as the throughput Si.j = SMD = SMD(Ptr .j fcc putation offloading of one or more applications. spent in transmitting the application data. where the index 0 stands for the local computation. (4) 0. The (3) posRAT (cx . . 2. we focus on (7) a scenario where each SMD request is restricted to a single In the case of a local computation at the jth SMD. i. . Applications Ptr j D Pid j O Ei. n. . cy ): the spatial position of the SMD. (2) ber of M RAT nodes. 2. . Pl .j BWi −k d i .j = + . . we propose to optimize the global energy App = App(O. with the respective features. Pl j O +yj (9) (6) posm d (cx . . This article has been accepted for inclusion in a future issue of this journal. M }. . for a generic SMD. since the general App has to evaluate the computation offloading benefit and se- case in which a user requests several applications at a given lect the most convenient RAT aiming at minimizing the energy time can be seen as aggregation of several single-application consumption of the user requesting the service. .j the distance between the ith RAT and the 0. Pl j O D. we will focus on the energy consumption. fm d . By defining xi. posRAT (cx . . Hence. it is   possible to write di. and after hav. Ei. and the amount of data to be exchanged. . as a binary variable influence the performance of the computation offloading. This simplification does In classical approaches. it is possible to express a generic application “selfish” or greedy in the sense that it does not take into account App as a function of O and D. . cy ): the spatial position of the RAT node. which is very (2) n: the number of devices connected to the RAT node. the generic RAT is a function depending on the above in idle—while the application is computed on the cloud server— features and the energy consumed for the local computation. Si. . the SMD computing an application not prevent the generalization of the system. for the local computation performed by SMDj . 1. if SMDj is assigned to RATi xi. if SMDj computes App locally yj = ing defined as di. . . In contrast to this. where the zeroth element represents the fictitious node related Each SMD is characterized by the following features which to the local computation. energy consumed by the jth SMD is composed by the energy where cx and cy represent the coordinates. Mobibyte [23]) because our study is mainly based on the communication links optimization. 2. .j = posRAT. where SNRj is a reference signal-to-noise ratio (SNR) ni = xi. we can write 1. .j = . This decision is requests. . D).j of the link between the ith RAT and the jth SMD. N. otherwise Given these entities.j = xi. where cx fm d j and cy represent the coordinates. 1. the energy spent by the jth SMD for offloading through the ith RAT is C. 2. . . where Thus.g. by exploiting (6). that stands for the presence of a communication link between (1) Ptr : the power consumed by the SMD to transmit data to the SMDj and the RATi . . . value at the receiver side and k is an attenuation factor due to l∈N the signal propagation Rearranging the constant terms and the variable terms. cy ) − posm d. as a binary variable accounting the RAT node. 2. assuming an environment where the devices of the set N = {1. cy ) (5) . . (3) consumption. N } are requesting to offload an D. N (8) fm d j application-focused studies (e. mance.j . j = 1.i (cx . We neglect other parameters usually considered in more E0. posm d (cx . Even if all the SMDs can simultaneously request the com. the energy consumed Thus. with the exception of pagination. i = 1. j = 1.

j (10) Si.j = log2 1 + SNRj e−k d i .j Ki.j id xi. BWi   tr Ei. j (6) ni l∈N .j = xi.l + Kj + yj E0.

j . 2. with the exception of pagination. fm d.j . D) Output: X = (xi. posRAT.n ← RATl . cy )) where the two constants are RATi .j ← 0∀i ∈ M∀j ∈ N yj ← 0∀j ∈ N Fig. Pl. SNRj . j } choose xi. .i (cx . Scheme of the biased-randomization approach.n + 1∀i ∈ M \ {0} tr Pt r j D calculateEi. the optimization problem can be expressed as . 2. .j = 1 and X ← xi. . This article has been accepted for inclusion in a future issue of this journal. M } // M number of RAT nodes N = {1.j Pid j O Y ← yj Kjid = . N } // N number of SMDs RATi = RAT(BWi . . 2.n − 1∀l ∈ M s. .t. 1.n ← 0∀i ∈ M \ {0} xi.j = min{El.j ) Y = (yj ) Initialization : RATi . Ei.n ← RATi . cy ))i ∈ M \ {0} App = App(O.j = BWi log 2{1+SNRj e −k d i . for j = 1 to N do SMDj = SMD(Ptr. xl. MAZZA et al. .j (cx . ) posm d. Content is final as presented.j . . ni . Pid. fcc end for Therefore.: SUPPORTING MOBILE CLOUD COMPUTING IN SMART CITIES VIA RANDOMIZED ALGORITHMS 5 Algorithm 1 Greedy Link Selection Heuristic Inputs: M = {0.j ∀i ∈ M Ki.j .j .j yj s.t. l ∈ M} RATl . .

In order to evaluate the effectiveness of our approach.j ∗ Ei. 1}. cy )) since we take into account mobile devices whose position can calculate di. that the allocation decision is post- the SMDs requests have been set. j∈N (11b) N = {1. posRAT.y i ∈M l ∈N j ∈N j ∈N Inputs: M = {0. . .j . we propose to use a fast probabilistic procedure. j xi . . we use an off-the-shelf commercial solver for j = 1 to N do able to provide optimal solutions. Thus. j . . 2. the objective function has to be solved repeatedly. Pl. it cannot be used for online allocation of resources. 1.j .j ascending ∀i ∈ M that the selection of the RAT by the SMDs has to be done in a choose xi. poned until all requests have been received—which in practice mization needs to know the position and the requests of all the might require customers to wait a few seconds for their requests SMDs. . min Z = xi . Moreover. instances.j . Ea.j ← 0∀i ∈ M∀j ∈ N or near-optimal solution to the previously described optimiza. fm d. First. However.. while ensuring that all devices are either assigned to a X = (xi. once all same time—or. X ← xi. Note that approach assumes that all users’ requests are processed at the the global optimum can only be found ex-post. OPTIMAL AND HEURISTIC SOLUTIONS d0. 2. i ∈ M) very short time. Ei.j ← d0 RATi . yj ← 0∀j ∈ N tion problem.j ∀i ∈ M change. M } // M number of RAT nodes s.j ) RAT or leaving the computations to be done locally (11b). . alternatively.j (cx .j . for real large-size SMDj = SMD(Ptr. j + yj = 1.e. it requires unsuitable computation times.t. i. This to be served.t.j . .j Our algorithm makes use of biased (oriented) randomization. xi . the performance is compared with the optimal solution. yj ∈ {0. . end for which represent a significant improvement over the previously proposed greedy algorithm [16] summarized below. j Kitr. (11c) App = App(O. ni . since the global opti. . Y = (yj ) Initialization : IV.n ← ni ∀i ∈ M \ {0} In this section.i (cx .j . SNRj .j yj s. j (11a) Algorithm 2 Biased-Randomization Algorithm x. Y ← yj and it allows us to obtain near-optimal solutions in real-time. Given sort di. posm d.j = geoinv(Ei. Pid. cy ))i ∈ M \ {0} xi . we present different methods to find an optimal xi. N } // N number of SMDs i ∈M RATi = RAT(BWi . i∈M j ∈ N. D) The objective function (11a) minimizes the overall energy Output: spent. l + Kji d + yj E0 .

Scenarios in case of 500 and 5000 SMDs. Thus.k for the SMDs already connected with the same A. but it also takes longer time than available to quickly while approaching the optimal solution in significant lower time serve SMD requests. reaching a solution closer to the optimal without tion. while each rent situation. modifies the values of the throughput Si. If the agent tends to maximize his/her individual utility function. The main idea behind the proposed difficulty for solving the QSAP comes from the fact that the approach is to introduce a slight modification in the greedy con- objective function is quadratic. antenna. this list is completed for each SMD. Hereby the heuristic logic is maintained but a heuristic methods. B. . otherwise. the interval analysis. This random effect as a reference to compare heuristic procedures (which can do the is generated by the use of a skewed probability distribution: at assignment dynamically or following a wait-and-go approach) each step of the constructive process. . Therefore. whereby the probability is higher for the more promising movements. promising” next node from the list of the available ones. j − 1. leading to a feasible but suboptimal solution. the optimal solver is only taken certain degree of randomness is introduced. i. respectively. the cost function is evaluated on the basis of the cur- This allows us to find a good solution in real time. . Content is final as presented. . This article has been accepted for inclusion in a future issue of this journal. the energy Ei. biased randomized method takes into account a solution from a the SMD will connect to the node which minimizes the cost collective or social point of view by simultaneously considering function. and one LTE eNodeB.k and. 3. Biased-Randomized Algorithm (BARON) [36].e. the heuristic method makes use of a “greedy” sorts each possible access node based on a self-calculated ob- behavior. Since the requests of the SMDs appear in time next step from a list of potential constructive movements [16]. In contrast. as shown in Section V. this . sequence. assigned to each potential movement. The computational than the optimal method. the link selection scheme allows us to select the “most needing to know all SMD positions. with the exception of pagination. consisting on the selection of the “most promising” jective function. it is important to note that if a uniform probabil- explained in [16].. which justifies the use of fast structive behavior. i. it will compute the application locally. Thus. reported The mathematical model (11) can be seen as a particular case in Algorithm 1. a set of users’ demands altogether instead of one user’s demand The selection of the ith node for connecting the jth SMD at a time. which ity distribution would be used instead of a skewed one. As Moreover. this approach out- position of all the SMDs requests before assigning them to an performs by far the solution provided by the greedy approach. for k = 1. and duality with advanced branch-and-bound proposed probabilistic algorithm is able to find near-optimal optimization concepts. hard [35]. connections. If the offloading in different dimensions when considering social or collective operation is advantageous with respect to the local computa.. quality of the solutions generated by the original heuristic tion problem following a greedy behavior [16]. Greedy Heuristic The use of biased-randomization significantly improves the A heuristic algorithm can be used to solve the link selec. three Wi-Fi APs. This solver combines constraint propagation. performance. 6 IEEE SYSTEMS JOURNAL Fig. solutions in “real time” a few seconds or even milliseconds Note that the QSAP solution not only needs to know the if parallel strategies are employed. does not take into account any forecast of future of the quadratic semiassignment problem (QSAP) which is NP. considering only the previous requests. Optimal Solution node. a selection probability is with an optimal solution from the performance point of view. consequently. The offloading cost is lower than the cost for the local computation. By exploiting biased-randomization techniques [15]. this strategy. 2. A first approach to find the solution is to use the global optimizer Branch-And-Reduce Optimization Navigator C.e.

Biased Rand. Biased Rand. solutions depending on some of the inputs of the model (i. whereas all elements are potentially eligible. throughput.30 13 745.3240 +2. Each time the randomized algorithm is executed. Dev.19 22 500. with the exception of pagination. MAZZA et al. [%] Avg.49 8239.67 7561. the cost function Ei. the biased-randomization of candidates—i. etc.74 TABLE IV in the sorted list.5161 +0. Dev.23 18 832.9327 -0.3169 ref.49 6345.50 1.0000 −2.14 0. which has been list—and then apply a uniform randomization to the order in also used for the greedy heuristic.4986 −0. Dev. [%] Local 22 500.70 Greedy 3181.26 0. 6343. a new prob- abilistic solution is obtained (see Fig. Dev.j . Dev.53 19347. is evaluated for every possible RAT node (also considering the ized algorithm—which can be encapsulated into a multistart fictitious node 0 related to the local computation). Instead of restricting the list of candidates.44 abilities are proportional to the position of each element in the Baron stopped after 1000 s list.70 9819. GRASP meta.22 TABLE II ENERGY [W*S] N SMDs 500 1000 2000 5000 Avg. That is. [%] Avg.1707 +0.31 14 716.e.5840 +0.49 35 438.1699 +0.3186 +11.1271 ref.e. allow us to improve those obtained through the greedy heuristic. This allows not N SMDs 500 1000 2000 5000 only to avoid the problem of selecting the proper size of the Greedy 0.53 8210. heuristics [37] propose the consideration of a restricted list Focusing on the selected problem. corresponding to those at the top of the based on the objective function defined in (11).3141 +11.6714 +14.7816 +0.8004 +1. which the elements of that restricted list are selected. Dev. [%] Avg.j of the device from the RAT nodes and on the ahead. elements at the top of the list have a TIME FOR COMPLETING THE ALGORITHMS [S] higher probability of being selected than those at the bottom. Dev.24 Baron 3.47 1.17 Nearest Node 4609. To avoid losing the logic behind the heuristic. 3.1570 −50.30 Baron 4104. Dev.53 6323.3240 +2.0000 +294. Dev. number of devices ni that potentially can be connected to each lection probabilities are assigned to each potential movement RAT node. 8236. 2). 0.6979 −14.85 1.3626 +14.6647 −10.73 19 711.42 Baron 3164. 3. 0.47 25 000.3416 ref.8480 ref. but we can also guarantee that the selection prob- Biased Rand.0000 +445.50 46 060. Biased Rand. Dev.8885 ref. the proposed approach allows us to achieve different random but not correctly oriented. In this When there is a new SMD request.2418 −15.88 TABLE III TIME FOR APPLICATION COMPUTATION [S] N SMDs 500 1000 2000 5000 Avg. [%] Avg.74 28.7847 −50.7438 +10. [%] Avg. 4106.26 22 500. different se.3306 −2. Some of these solutions improvement would rarely occur since the logic behind the con. The obtained process—while most of the logic or common sense behind the values are sorted in such a way that the probabilities of being original heuristic is still respected.3241 +100. Dev.20 0.9261 ref.65 4.6292 −0. 16 447.46 Greedy 3. selected depend not only on the cost function Ei.6194 +19.4181 −0. 21 007.5693 −50.0653 ref. [%] Avg.20 14 494.41 0.7909 ref.0000 +173. 3166. 1.).5039 ref. 12 659.7999 +1.61 20 918.69 25 000. structive heuristic would be destroyed and the process would be Moreover.64 11.3926 −50.9989 −0. latency.7944 −0. [%] Avg.. This article has been accepted for inclusion in a future issue of this journal..j way.5844 +0.04 14 848.0000 +19.29 0. a deterministic procedure is transformed into a random. Content is final as presented.9142 ref. [%] Local 25 000. . but also on The proposed biased-randomization approach goes one step the distances di.57 20 432.5799 ref.26 0. [%] Avg.04 14 003.13 25 000. 22 999. a sublist including just some of the most algorithm for the link selection consists in a skewed criteria promising movements.26 Greedy 4126.1551 ref. [%] Avg.36 0.33 0. energy.2457 +19.34 6.0000 +685. [%] Nearest Node 1.01 Nearest Node 3553.80 22 500.: SUPPORTING MOBILE CLOUD COMPUTING IN SMART CITIES VIA RANDOMIZED ALGORITHMS 7 TABLE I THROUGHPUT [MBPS] N SMDs 500 1000 2000 5000 Avg.0152 +8.4755 −0.7108 +68.07 restricted list.0000 +97. Dev.0000 +36.2169 −9.

the capacity constraints on the antennas is not considered. higher improvement in terms of energy is expected. and SNR equal to 30 dB. This article has been accepted for inclusion in a future issue of this journal. Pid equal to 0. However. throughput.9 W. all the SMDs compute lo. application computation in Table III. Note that the modifications included in Algorithm 2 and the time employed for accomplishing the application are do not increase the computational complexity with respect to evaluated and observed for each configuration. We considered a HP iPAQ PDA with a consumption. V. the results related to the In this section. since the The parameters fm d . 3 represents the by the exact solver—in much lower computation times.e. while it is more related to We have chosen an application which is accomplished through the ability of the algorithms to solve the problem in a fast way a number of operation O equal to 107 and. A deployment tion has been computed with Baron under the Neos server area equal to 1000 × 1000 m2 is considered. The On the other hand.0). 1000. (c) Throughput for 5000 SMDs. the energy. For the cloud server.24% in throughput. 400-MHz Intel XScale processor (fm d equal to 400 MHz) and Table IV shows indeed the time needed by the algorithms the following values: Pl equal to 0. the one hand. respectively. Content is final as presented. On data transfer D equal to 104 bits. which has been stopped after 1000 s of computational chosen a controlled random number generation of the Mersenne time.3 W. with the greedy heuristic. the energy consumed by the SMDs. 8 IEEE SYSTEMS JOURNAL Fig. scenarios are considered. the average eNodeB/AP. Furthermore. randomized algorithm is below a few seconds.j · Ei. the criteria for sorting the SMDs characteristics) are also taken into account for compar- choice list takes into account the product di. equal to 1.999) and The results show that the biased-randomized algorithm (1000. similarly to [16]. we choose to compare all the values instead of Ei. This demonstrates that our probabilistic algorithm is able Twister type and a seed equal to 1 (this pseudorandom number to provide near-optimal solutions—similar to the ones provided generator is implemented in MATLAB). we have Baron. An attenuation coefficient k for the propagation equal clearly outperforms the link selection configuration obtained to 10−3 has been considered. computation times that cannot be employed in real-life practice. hence. Comparison of the performance in terms of throughput for different SMDs number.neos-server. each of them involving an increasing number of SMDs: 500.3.e.500) and three Wi-Fi access points (AP) with a bandwidth ods were implemented in Matlab.j is adjusted empirically and with those obtained with the greedy algorithm as reference.j . equal to 22 MHz are positioned at point (0. 2000. NUMERICAL RESULTS Tables I–III show. Also.30% in energy consumption. The model was implemented with eNodeB with a bandwidth of 100 MHz is positioned at point the mathematical modeling language GAMS. The ex-post solu- pare the performance of the different algorithms. rithm 2.0). (b) Throughput for 2000 SMDs. where one LTE (http://www. and the time. which is still a cally) and the nearest-node configuration (i. ergy. ison purposes. The value of d0. In particular. Ptr for achieving a stable solution of the link selection problem. and reductions of 14. (500.. (a) Throughput for 1000 SMDs. Pid . and 5000. This strategy is reported in Algo. Fig. The throughput. This connected to the closest RAT independently from the RATs and is due to the fact that the biased-randomized algorithm is run . 4.j . a realistic test case is evaluated in order to com. if offloaded. Pl . Ptr . The region assuming 500 and 5000 SMDs. needs a and. the computing time employed by the biased- no-connection configuration (i. The other meth- (500. respectively. en. such time is not directly related to the time needed for the we suppose a computation speed fcs equal to 106 MHz [20]. Moreover. Improvement of 2. To improve the link selection. and SN R depend on the objective function in (11a) is designed for minimizing the energy selected mobile device. and 0. each SMD is little higher than the one employed by the greedy algorithm. approach. as. energy and time values provided by the randomized algorithm The location coordinates of each SMD in the deployment area are similar to the optimal/near-optimal solutions provided by are sampled from a uniform distribution. due depends on the overall number of SMDs which are expected to its strict relationship with the proposed biased-randomized to interact in the system. and time needed for accomplishing the application. different Algorithm 1.. with i ∈ M. it has to be noted that the exact solver requires The algorithms are compared in terms of its exploitation in a real-life dynamic scenario.42% suming there is no limitation on the number of SMD per each in computational time can be observed. with the exception of pagination.

MAZZA et al. energy consumption. 5. Performance in terms of throughput. . (c) Biased-randomized configuration. 6. Link selection in case of 500 SMDs. This article has been accepted for inclusion in a future issue of this journal. (c) Latency. (a) Nearest node configuration. Each gray indicates that the SMD is connected to the RAT marked with the same gray. and latency for a variable number of devices. (a) Throughput. (b) Energy.: SUPPORTING MOBILE CLOUD COMPUTING IN SMART CITIES VIA RANDOMIZED ALGORITHMS 9 Fig. with the exception of pagination. Fig. (b) Greedy heuristic configuration. (d) Ex-post optimal configuration. Content is final as presented.

7. (d) Ex-post optimal configuration.e. must be updated very often. 5(b) shows that the biased- randomized algorithm could be easily parallelized—by simply randomized algorithm clearly outperforms the greedy heuristic. (b) Greedy heuristic configuration. (c) Biased-randomized configuration. com. 10 IEEE SYSTEMS JOURNAL Fig. Content is final as presented. the average throughput values are shown Figs. to minimize the overall energy consumption. Moreover. to perform the application.. This means that the capacity of the for example. Thus. Comparing these . with the exception of pagination. it is still an Considering the average energy consumed by one SMD acceptable computing time. SMDs equal to 500 and 5000. Link selection in case of 5000 SMDs. Nevertheless. The greedy heuristic offloading applications which need a fast link selection decision. while providing solutions that tend algorithm is very similar in terms of average throughput. This article has been accepted for inclusion in a future issue of this journal. running each iteration of the multistart process in a different The same observation can be inferred from Fig. 6 and 7 report the different configurations (of the de- in a different scale to make possible the comparison of the vices placed in the observed area) for an overall number of three algorithms. This is useful for pared to Baron’s ex-post configuration. 4. respectively. respective results of Baron. when the users are moving and the configuration RAT nodes is well exploited in the biased-randomized algo. the biased-randomized al- The results are visually compared in Figs. 4 and 5(a). is clearly outperformed. Each gray indicates that the SMD is connected to the RAT marked with the same gray. where gorithm can be exploited by the SMDs to reach near-optimal we can see that the performance of the biased-randomized configurations in real time. 5(c) for the core or computer—so its computing time would be the same as average time that a device needs to accomplish the application. (a) Nearest node configuration. the one employed by the greedy algorithm—i. if necessary the biased. Fig. In Fig. in the order of Both the average energy and time values are very close to the milliseconds. rithm. several times in a multistart process. whereas the black SMDs are computing the application locally.

. 11. vol. no. Ratti. 45–55. no. no. no.” IEEE Wireless Commun. 2015. Appl. pp. 2011. Available: http://esa. since it aims to save SMD’s en. vol. “On the use of Monte Carlo simulation. Ur et al.html Res. Commun.” IEEE Trans.” Int. [34] A. no. Huang. Pinedo.” IEEE algorithms for the arc routing problem. “A survey of mobile cloud [32] A. vol. 51–56. U. we can note that the case referring to the biased. 2012.. vol. no. in these figures. H. pp. improving performance and floading computation save energy?” Computer. A. no. Madani. 11. York.” in Proc. REFERENCES pp. N. pp. and G.. the mobile communication system is managed in [19] F..” J. “QoS-aware energy-efficient resource allocation in OFDM. 3rd ed. Res. 18. “A dynamic offloading algorithm the field of smart cities and mobile telecommunications. TX. NY. 47. 70–71. Wang. Kumar and Y. Cisco. pp.-G. 821–834. Wang. V. the Clarke and Wright savings heuristics.” Eur. P. applications. [20] K. R. vol. no. Maniatis. M. “An annotated bibliography of GRASP— [4] Scientific American. Resende. M. 2011. “A user-satisfaction based of- VI.1002/dac. Liu et al. 2. Some for mobile computing. Mag. “Communicating while performing local computation. 28. Juan. Algorithms. Serv. Furthermore. vol. 2012. vol. Oper. S. Jan.” Comput.. 2015. Z. Thus. 2013.” AI Commun. Chen. Khan. González-Martı́n. Hui. Y. vol. Masip. T. 6th Conf.. T(2014)507480_EN. Mar. vol. Han.. cation. n-job flow-shop sequencing problem. Di Lorenzo. Designing the Urban Future: Smart Cities. and S. 42–49..” IEEE Trans. no. 2007. pp. 2013. 3. pp. cache and splitting techniques to improve approach.. “Cloud computing for mobile users: Can of- sion of services in a smart city. D. M.. “Development and assessment of the SHARP and RandSHARP [6] M.europa. allows to adapt the link selection based on the node density. 3. Lourenco. Ham. May. Grid Comput. R. Chun. and G. and B. considering the QoS of the entire population Architectures.Euro capacitated centered clustering problem. 34–44. Sardellitti. and smart world. [13] S. Chen. 2783–2788. efficient link selection in pervasive wireless environments. no. Jun. Content is final as presented. doi:10. Ihm. 103–126. connected to each eNodeB/AP of the same gray or. Castellà. Faulin. N. N. Syst. well-being of the community through a social use of HetNets Apr. 13. Brimberg. 552–558. Available: http://www. Riera. pp. pp. E. pp. Nov. Nov. 2014. This vision could be adopted for the provi.. 2151–2162. 2. no. 6. Meng. and A. Srinivasan.” in Proc. [27] E. Res. no.” Wireless Com. 173–189. Naik. Marathe. Zheng. Chaves and L. Tarchi. 2015. and P. 301–314. Pérez. 2013. “A survey for the quadratic assignment problem. MA. pp. S. Querido. Cambridge. “A heuristic algorithm for the [9] R.. A. E. Juan. Dinh. and [1] World Urbanization Prospects. Festa and M. Mar. no. 2007. Nawaz. Barrios. Tuts.. “Green Internet of Things for [31] S. parl.” European Par. 2011. Lorena. parametrization. Jorba. J. vol. 2015. S.. 393–413. “Context-aware mobile cloud comput- [10] L. vol.. Soc.” IEEE Trans. vol. 2. C. we have presented different strategies for an [17] W. Available: http://www. 657–690.. and A. Mobile Syst. vol. NY. Enscore Jr. In vol. 1983. 14–22. a greedy or individual one—during the real-time resource allo. Castella. O. 1587–1611. U. [16] D.. Jan. 1991–1995.. In particular. In this paper. T. Wen. 3. 8–9. P. [Online]. with the exception of pagination... 2009. Boaventura-Netto. USA: Springer. p-median problem: A survey of metaheuristic approaches. Parallel Distrib. Dec. USA. J. Caceres-Cruz. Dohler. Lei.. 6. Rep. 3. V. USA: Scientific Amer. Conf. the nodes represent SMDs mobile cloud computing application models. Commun.-H. Mag. and Systems. D. 26. Jan. 18.. [21] E. no. 31. Niyato. and P. Oper. challenges. vol. [Online]. ergy. numerical results contribute to illustrate the potential of the pp. “Us- computing: Architecture. Res. 2008. no. D. Hansen. 20.” IEEE Cloud Comput. Ur. Shao. Jun. Othman. 2014. de Abreu. 13. [14] X. and approaches. pp.-H. Corazza. solution. Wireless Commun. 2014. 6.” Int. from an ecofriendly perspective. L. Niyato.. [29] A. New York. pp. “MAUI: Making smartphones last longer with code and MCC. M. [7] H. A. 49–62. S. 1085–1097. it can be seen as an enhancing method offload. 2013.” Eur. no. CA. pp. C. ing iterated local search for solving the flow-shop problem: Paralleliza- mun. A. A. Zhong. 2. 2. [23] A. vol. A. J. and H. Austin. Syst. Oper.” J. optimization behavior of the proposed biased-randomized [15] A. The New Science of Cities. 5. Riera. “CloneCloud: Elastic execution between mobile device and cloud.” IEEE Access. Zhang. 20. no. R. 179. 1. Mladenović. and Q. Q. Trans. A. of mobile users. It is computing: Distributed mobile cloud computing over 5G heterogeneous possible also to highlight that the biased-randomized approach networks. “The 2016). which have been used in the past to solve similar Comput. and I.. 2015. D. Res. 2012. Loiola. Manville et al.. M. vol. 2015.. A. Y. vol. R. S. 4. “MobiByte: An application development model for mobile transportation. 2014. Res. vol. vol. 2014. 2014. “Mapping smart cities in the EU. J. Juan. G. “Gearing resource-poor mobile devices with powerful clouds: a cooperative way. Population Division. Qian. D.” IEEE Signal Process. for increasing cellular capacity and coverage. Barbarossa. IEEE Global Commun. 37. alternatively. Falconer. M. Lee. Mazza. 6. represented as black dots. pp. Kota. 1. New Part II: Applications.un.2931 . 2010. Luo. no. USA: MIT Press. pp. Oper. J. Hu. pp. J. Ngai. Juan. combinatorial optimization problems in the fields of logistics. and H. P. 4. “Challenges on wire- randomized algorithm is very similar to the ex-post optimal less heterogeneous networks for mobile cloud computing. Guimarans. allows to consider a more global point of view—instead of just “Mobile data offloading through opportunistic communications and social participation. Leung. 2013. Hahn.. P. we have proposed a novel randomized algorithm that [18] B. Apr. Apr. United Nations. Tech. “Toward transcoding as a service: energy-efficient offloading policy for green mobile cloud. 4. E. 1–32. Jun. P. Our approach is based on the use of biased-randomization [22] B. [3] C. pp. service-provider/visual-networking-index-vni/index. “Clustering search algorithm for the liament. no. San Francisco. Kumar. [2] Cloud and Mobile Network Traffic Forecast. Shu. no. Jun. A. [Online]. Syst. 67–73. tion. D. [33] M. Oper. vol. 21. (2014). pp.” IEEE Commun.” in Proc. J.: SUPPORTING MOBILE CLOUD COMPUTING IN SMART CITIES VIA RANDOMIZED ALGORITHMS 11 configurations. Commun. Muñoz. K. and A. techniques. 91–95.” IEEE Wirel. no. 927–939. 51. 43. Batty. Riera. vol. Chen. Khan. (Jun. This article has been accepted for inclusion in a future issue of this journal. Ed. Jun. Scheduling: Theory. and D. [12] A. and production. ing and its challenges. vol. Paraszczak. 16.. Moreno-Pérez. 11. V. and E. 974– density in a limited way. pp. Conf. 176. 8th Int. pp. 3. [11] L./Jun.” Int. VNI. 2010. M.. Patti. Austria. based heterogenous cellular networks.. Jan. where the SMDs connected to the same RAT are clus. [24] D. 62. [25] M. Lu. MAZZA et al. 605–628. Surv. [28] N. J.” IEEE Netw. This confirms the “socially aware” 983. pp. “Rich vehicle routing problem: Survey.pdf [30] P. Surv. A. 25. Arias.. 16. and A. Trans. Jun. pp. Mobile Comput. CONCLUSION floading technique for smart city applications. Mateo. vol. P. Apr. 131–172. Khan.. pp. 2014. 1. Oper.. This work extends their use to cloud computing. Mobile Comput. no. and G. proposed approach.” Omega. particular. Giambene. Zhu. Cuervo et al. 6. C. Zhou et al. Dec. R.. pp. Khan. 3. and applications. [26] J. pp.” IEEE Wirel. P. and randomization issues. 2010. Commun.. [5] C. J. USA. L. [8] B. Jun. A. and J. “Decentralized computation offloading game for mobile cloud while the greedy approaches only take into account the nodes computing. “A survey of tered. 3.” ACM Comput. “Smart cities. Salzburg. “Hetnets—A new paradigm m-machine.

Norway. Italy. His research inter. . versity of Bologna. P. the IEEE 2009 Satellite Communications Distinguished Service Award.. pp. Italy. Glob. Content is final as presented. She is an expert on system optimization in pub- lic administration organizational contexts at Emilia Romagna Local Government. Italy. Internet of Things. trial engineering.D. Angel Alejandro Juan received the M. Tarchi has been an Editorial Board Member for the Wiley Wireless Com- munication and Mobile Computing. a and has worked as a Project Manager for technological packaging industries Member of the Alma Mater Board of Directors. Spain. Integral Satcom Initiative. degree in informatics and telecommunica. He nic University of Catalonia (UPC). He was the General Chairman of the IEEE ISSSTA 2008. was Head of the Department of Electronics. the Chairman of the School Department of Computer Science. 6. telecommunications. “ Quadratic assignment problems. He was an In- vited Researcher with the University of Southampton Daniela Mazza received the master’s degrees in elec.). the Founder of the Marconi Institute for Creativity (2011). Open University of for Telecommunications in 2000–2003. Tawarmalani and N. Society Best System Paper Award. and a postdoc- pp. Her re. D. Bologna. the Founder and Chairman of the Researcher at UPC and Norwegian University of Sci. 1998. Spain. degree in indus- New York. 2013. Barcelona. She was a Postdoctoral Mobile Systems Task Force (ASMS TF). pp. the 2nd International Symposium on. respectively. M. 12 IEEE SYSTEMS JOURNAL [35] R. Cambridge. degree in binatorial Optimization. founder of the Mavigex S. 1995.K. and R. University of Bologna. Cambridge. ASMS 2004–2012 ests include the telecommunication area. Wireless Communication Systems 2005. He is currently an Asso- procedures. and energy systems. He was the recipient of the Marconi International Fellowship Young Scientist Award Daniele Tarchi (S’99–M’05–SM’12) received the in 1995.D. mation and synchronization. Prof. Jun. Trondheim. Italy. Eds. spread spectrum and multicarrier transmission. navigation and positioning. G. tional Symposium on Spread Spectrum Techniques and Application (ISSSTA) Florence. Italy. University. (U. 103. NY. Bologna. Pardalos. at the IEEE International Conference on Telecommunication 2001.” in Handbook of Com. at LAAS-CNRS (France). He has authored University of Bologna. a Member of the Board of the 5G Infras- tructure Association. Dr. Com- She is currently a Postdoctoral Researcher at the puter Science and Systems (DEIS) in 2009–2012. with the exception of pagination. He has been involved in several national projects as well as European projects and has been active in several industry funded projects. in 2000 and 2004. degree in telecommunications engineering and the 2013 Newcom# Best Paper Award. 2741–2814. applied mathematics and the Ph. the Best Paper Award of the IEEE Interna- tions engineering from the University of Florence. V. with partic.D. and scientific creative thinking. production.-Z. 2. the 2002 IEEE Vehicular Technology the Ph. and the statistics and operations research from the Polytech. Du. His research the Ph. a Member of the Marconi Society Board of Directors. including solution for distributed multimedia system and ser- Giovanni Emanuele Corazza (M’92–SM’07) is cur- vices management in a context of smart city.. Barcelona. 2005. Bologna. Mar. Program. Graham. toral internship at the MIT Center for Transportation [37] T. ciate Professor of optimization and simulation with the Department of Computer Science. 109–133. communications. no. C. and MIC Conference 2013.l.D. Optim. USA. vol. This article has been accepted for inclusion in a future issue of this journal. He has authored or co-authored more than 260 papers. since 1991. the Vice-Chairman of the Net- Adela Pagès-Bernaus received the Ph. USA: Springer. esti- computing. Burkard. interests include applications of randomized algorithms and simheuristics in lo- and information technologies engineering from the gistics and transportation. or co-authored more than 150 peer-reviewed papers in these fields. Internet channel characterization. degree in World2020 European Technology Platform. “A polyhedral branch-and-cut He completed a predoctoral internship at Harvard approach to global optimization. “Greedy randomized adaptive search and Logistics. She pos. L. vol. at the University of Natural Resources and tronic engineering and communication science and Life Sciences (Austria). and Internet computing. He was the Symposium Chair at IEEE Wireless Communications and Networking Conference 2011 and at IEEE Globecom 2014. His search interests include operations research applied research interests include wireless and satellite communications. ular interests to resource allocation and link adapta- tion algorithms in wireless and satellite networks. Barcelona. the Chairman of the Advanced Satellite Catalonia.Sc. Resende. A. 2. degree in electronics.). and the Workshop Co-Chair at the IEEE International Conference on Communications 2015. rently a Full Professor with the Alma Mater Stu- sesses more than 15 years of experience in the local government organizations diorum. Hindawi Journal of Engineering. M. E. Feo and M. Conferences. 225–249. Sahinidis. and the Scientific World Journal since 2012 and has served as an Associate Editor for the IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS from 2008 to 2013.” Math.K. a European technology platform devoted to satellite ence and Technology. Corazza served as an Editor for Communication Theory and Spread Spectrum for the IEEE TRANSACTIONS ON COMMUNICATIONS in 1997–2012.Sc. and at He is currently an Assistant professor at the Uni. MA. Spain. no. Open Univer- sity of Catalonia. and at the University of Portsmouth (U.r. spin-off company.” J. Her research in- terests include the services to the citizens. [36] M. mobile radio to the fields of logistics and transportation.