Throughput Optimal Random Medium Access Control for Relay Networks with TimeVarying Channels
Abstract
The use of existing network devices as relays has a potential to improve the overall network performance. In this work, we consider a twohop wireless relay setting, where the channels between the source and relay nodes to the destination node are time varying. The relay nodes are able to overhear the transmissions of the source node which may have a weak connection to the destination, and they help the source node by forwarding its messages to the destination on its behalf, whenever this is needed. We develop a distributed scheme for relay selection and channel access that is suitable for timevarying channels, and prove that this scheme is throughput optimal. We obtain the achievable rate region of our proposed scheme analytically for a relay network with a single source and a single relay node. Meanwhile, for a more general network with more than one relay nodes, we perform MonteCarlo simulations to obtain the achievable rate region. In both cases, we demonstrate that the achievable rate region attained with our distributed scheme is the same as the one attained with centralized optimal scheme.
1Introduction
1.1Motivation
In wireless networks, there is always a node which has the worst channel quality (e.g., cell edge node) due to its location, shadowing effects or physical radio capabilities. Wireless channel quality for these nodes may become so weak that in order to maintain a reliable communication, some relays have to be employed on the path between the source and the destination node to aid their communication [1]. Instead of deploying separate devices solely used for relaying, it is possible to utilize the existing wireless devices in the network for the same purpose. It is expected that future wireless devices are going to be highly adaptive and flexible, and thus, opening up new paradigms where they can adapt themselves utilizing any network resource, e.g., as other peer nodes in their vicinity, to improve their communications.
One of such paradigms is relay networks where the aim is to improve the network throughput by allowing wireless nodes to participate in the transmission, when they are neither the intended destination nor the source [2]. A widely used wireless relay network model consists of two hops, i.e., there are a source, a destination, and multiple relay nodes [3] as in Figure 1. The basic idea is that the relay nodes, overhearing the message transmitted from the source node, forward the same message to the destination node instead of treating it as interference [4].
Fundamentally, this idea of relaying is very similar to routing in multihop networks, and there is already a rich literature on the subject. In particular, centralized throughput optimal policies were developed for opportunistic multihop relay networks in [5]. A policy is throughput optimal in the sense of having an achievable rate region that coincides with the network achievable rate region, and thus, a superset of the achievable rate region of any other policy. However, a centralized policy requires global information of the entire network. The collection of this information is costly in terms of power consumption and time. On the other hand, distributed (decentralized) policies may alleviate these implementation costs by running a loosely coordinated scheme at each network node.
A common network setting is where the source and relay nodes are located in such a way that the source is furthest to the destination, while the relays are relatively closer. This notion of distance includes both real distances in freespace communication, or virtual distances that also include the effects of fading and shadowing [8]. Hence, the source has the worst channel quality in terms of probability of successfully transmitting its packets to the destination.
There is a number of works developing distributed policies for twohop wireless relay networks with the objective of improving the transmission rate of the source node [9]. In these works, the timevarying wireless channel is modeled as an ONOFF random process for analytical tractability, where a packet is successfully decoded at the destination node when the channel is ON or it is lost when the channel is OFF. Note that modeling the channel as an ONOFF random process is an approximation of a more general timevarying channel model, where the channel is a multistate process, with each state supporting a maximum transmission rate with which the data can be reliably sent [13]. The authors then propose suboptimal channel access algorithms, and calculate the achievable transmission rate of the source node by employing Loynes’ theorem [14].
Unlike prior works, we develop a throughput optimal distributed policy called Relay QCSMA (RQCSMA), for relay networks with timevarying channels, and show that its achievable rate region coincides with the achievable rate region of a centralized optimal policy. Our policy achieves this without the knowledge of the channel statistics, and it performs the scheduling of the source and relay nodes by only utilizing local information without any explicit message exchanges.
1.2Related Work
The network setting used in this paper was previously studied under different contexts and assumptions in the literature. In particular, for multihop opportunistic routing it is shown that Maximum Weight Scheduling (MWS) algorithm, which is a centralized scheme is throughput optimal [5]. MWS and many of its variants exist in the literature addressing the optimal resource allocation under different network assumptions and for different applications, e.g., cognitive networks [15], relay networks [16], etc [17]. Unfortunately, MWS is known to be NPhard for general networks [19] and it is not amenable to a distributed implementation.
The high time complexity of the centralized algorithms such as MWS was addressed in the literature by developing lowcomplexity suboptimal algorithms. In particular, maximal scheduling is a lowcomplexity alternative to MWS that is amenable to parallel and distributed implementation [21]. However, maximal scheduling may only achieve a fraction of the achievable rate region [22]. Greedy Maximal Scheduling (GMS), also known as LongestQueueFirst (LQF), is another lowcomplexity alternative to MWS [23] with a complexity that grows linearly with the total number of the links [24]. Its performance has been observed to be close to optimum in a variety of wireless network scenarios [25]. Although these algorithms reduce the complexity of MWS, they have two main shortcomings. First, they require network wide queue length information exchange, and second, they are mainly suitable for links with timeinvariant properties.
Another class of distributed scheduling policies, called Queuelengthbased Random Access Scheduling policies, use local message exchanges to resolve the contention problem [26]. By adjusting each link’s contention probability using the link’s local queue information, it provides explicit tradeoffs between efficiency, complexity, and the contention period. CarrierSensingMultipleAccess (CSMA)based scheduling policies [28], reduce the complexity by simplifying the comparison process, by exploiting carriersensing. Nonetheless, these results indicate that good throughput performance may be attained for time invariant channels using algorithms with very low complexity.
In practice, however, most wireless systems have time varying characteristics. When link states vary over time, the system throughput can be improved by scheduling those links with better states. This is known as the opportunistic gain [30]. However, many of the lowcomplexity scheduling algorithms in the literature do not exploit the opportunistic gain, and their performance in time varying channels are shown to be significantly lower [31].
Recently, a few other lowcomplexity schemes have been proposed that are provably efficient over time variable channels [31]. However, these algorithms, depending on the network structure, require either local or global queue length and channel state information (CSI) exchange. In Opportunistic ALOHA [34] the transmission probability is allowed to be a function of the channel state information and maximum throughput of the system is achieved. Distributed Opportunistic Scheduling (DOS) [36] involves a process of joint channel probing and distributed scheduling. The authors show that the optimal scheme for DOS turns out to be a pure threshold policy, where the rate threshold can be obtained by solving a fixedpoint equation. However both Opportunistic ALOHA and DOS are designed under the assumption of saturated queues and their performance have no guarantees for unsaturated queues.
The use of relay nodes has also been investigated in the context of cellular networks. In [38] a throughput optimal centralized downlink scheduling policy is developed for multihop relaying in a cellular network. Note that the uplink scheduling is more challenging, since the queue length information is not readily available at the base station. Conventional cellular architecture treats the mobile users as simple transceivers and due to this rigid clientserver scheme, they are fully commanded by cellular base stations. In future cellular architectures, mobile users will be given more freedom, and thus, they will be treated as local microoperators to improve network coverage [39]. DevicetoDevice (D2D) communication is another paradigm which appears to be a promising component in the next generation cellular networks [40]. D2D communication was proposed in [41] to enable multihop relaying in cellular networks. The potential of D2D communications for improving spectral efficiency of cellular networks was also identified in [42]. There are numerous distributed D2D policies [44], addressing the data relay services enabled by existing devices in the network. However, most prior works in the literature ignore the network stability and do not investigate throughput optimality.
1.3Main results and organization
In this work, we take both channel state and queue length information into account when scheduling the source and relay nodes. We propose a distributed CSMA based scheduling algorithm which is analytically shown to be the throughput optimal. The time complexity of the algorithm is shown to be linear in the number of nodes. The rest of the paper is organized as follows. In Section 2, we introduce the relay network model. In Section 3, we give a brief introduction to MWS and later propose our scheduling and contention resolution algorithms. In Section 4, we analytically prove throughput optimality of our algorithm. We establish the achievable rate region for a source and a single relay node in Section 5 and finally, we evaluate the performance of different scheduling algorithms via simulations in Section 6.
2System Model
In this section, we introduce the details of our system model and the definitions of the main concepts that are used in relation to this model. Throughout the paper, the following notations are adopted. Upper case and lower case bold symbols denote matrices and vectors, respectively. Upper case calligraphic symbols denote sets.
2.1Network Model
We consider a wireless relay network consisting of a source node and relay nodes all wishing to communicate with a common destination node as shown in Figure 2. We consider a timeslotted system where the time slot is the resource to be shared among different nodes. We adopt a noninterference model where only one node is transmitting at a given time
We assume that the channels between nodes and the destination node are timevarying ONOFF channels, where a packet is either successfully decoded at the destination or lost with a certain probability. The nodes with better channel conditions to the destination have higher successful transmission probabilities. We will refer to the source node as node and relay nodes as node , . We assume that relay nodes are in close proximity to node so that the channels between node and nodes are always ON and node can always transmit successfully to any node . This assumption is realistic in practical scenarios (e.g., see [48]), due to proximity of source node and relays.
Let be the channel state between node , and the destination at time slot . The random channel states, , are assumed to be independent and identically distributed (iid) across time and nodes. Let be the probability that the channel between node and the destination is ON, i.e.,
Let be the number of packets arriving to node at time . The arrival processes are assumed to be iid across time and nodes. Let be the rate of arrival to node . We also assume that and is finite for all and . The incoming packets are stored in a queue until they are finally transmitted.
Let be the size of the queue storing the packets of node at time . Meanwhile, each node also keeps a separate queue for node packets that they are supposed to relay on its behalf. Let , be the size of this queue at time .
Finally, we assume that the transmission rate of each node is the same and equal to one unit sized packet per time slot.
2.2The Relaying Scheme
At the beginning of time slot , the channel is acquired by one of the nodes according to the scheme to be explained in the subsequent section. If node acquires the channel, it sends its packet to the destination if , i.e., the channel between node and destination is ON. Otherwise, it forwards its packet to one of the nodes for future delivery of the packet to the destination via the selected node , . Node forwards the packet to node , where . Meanwhile, if a node other than node acquires the channel in a time slot, it may transmit either one of its own packets or a packet received from node in previous time slots. Node , transmits a packet from one of its two queues, i.e., or , with the largest backlog. The destination sends an Acknowledgment (ACK) message after the reception of the packet. ACK message is perfectly received by both node and the transmitting node , , and thus, node is cognizant of the delivery of its packet by the node , . Note that the assumption of perfect reception of ACK/NACK is not necessary. It can be shown in a similar way as was done in [13] that as long as the queues are updated periodically with period , our results still hold.
Let be the scalar denoting the scheduled node for transmission, where denotes the Idle state indicating no nodes are scheduled for transmission. Based on the aforementioned relaying scheme, the queues evolve as follows:
where is the indicator function, and its value is equal to if the condition inside the bracket is true, and otherwise.
3Distributed Algorithm
3.1Background and Preliminaries
Before we proceed with the discussion of our proposed distributed scheduling algorithm, we give a brief overview of the wellknown Maximum Weight Scheduling (MWS) [18] and QCSMA algorithms [29] in the context of our relay network.
Maximum Weight Scheduling
According to MWS algorithm, each node is assigned a weight that is the product of its differential queue length and the channel state. MWS algorithm chooses the node with the maximum weight for transmission at time slot . The differential queue length of the node is the difference of the queue lengths of the ingress and egress nodes. Note that in our relay network, the destination is the sink node, and thus, its queue length is always zero. Let be the maximum differential backlog of node with respect to nodes at time slot , which is defined as
In our network model, we assign weights to nodes in the network, and aim to schedule the node (and the queue of that node) with the highest weight at every slot. The weight of node at slot , , is defined as follows:
where , are functions that should satisfy the following conditions [13]:
is a non decreasing, continuous function with .
Given any , and , there exists a , such that for all and , we have
It can be shown that for this system model, the centralized MWS algorithm is throughput optimal using the Lyapunov drift theorem [18]. However, note that MWS requires the complete instantaneous queue length and channel state information.
Queue based CarrierSensingMultipleAccess (QCSMA)
CarrierSensingMultipleAccess (CSMA)based scheduling policies [28], reduce the complexity of MWS by simplifying the comparison of the weights exploiting the carriersensing mechanism. In QCSMA, each time slot is divided into a contention and a data slot. For any node , let be the set of conflicting nodes (called conflict set) of node , i.e., is the set of nodes such that if any one of them is active, then node cannot be active. In the contention slot, the network first selects a set of nodes that do not conflict with each other, denoted by . Note that this set of nodes is also a feasible schedule, but this is not the schedule to be used for data transmission. is called the decision schedule at time .
The network selects a decision schedule according to a randomized procedure, i.e., it selects with probability , where , and is the set of all possible feasible schedules of the network. Then, the scheduling procedure proceeds as follows: Each node within is checked to decide whether it will be included in the transmission schedule . For any if no nodes in were active in the previous data slot, i.e., , then chooses to be active with probability and inactive with probability in the current data slot. If at least one node in was active in the previous data slot, i.e., , then will be inactive in the current data slot. Any node maintains its state (active or inactive) from the previous data slot.
3.2Relay QCSMA Algorithm (RQCSMA)
We propose RQCSMA algorithm to obtain a feasible throughput optimal schedule, when the channels to the destination are randomly varying over time. In RQCSMA, a time slot is divided into a contention and a transmission slot as in QCSMA. Unlike QCSMA, the purpose of the contention slot in RQCSMA is not only to generate a collision free schedule to be used in the transmission slot but also to infer the channel states of every node, i.e., .
Let us denote by , the decision schedule at time slot with , where , and is a realization of randomly varying channel states. Note that in the relay network considered, the cardinality of set is always one, since only one node can be scheduled without conflicting with others. The network selects a decision schedule randomly with probability , where , and is the set of all possible feasible schedules of the network when . Note that in the relay network considered, a schedule is feasible, if only one node is scheduled, and the scheduled node has its channel in ON state, i.e., . Meanwhile, it is always feasible to schedule node , since links between node and other nodes are always ON.
Unlike QCSMA, in RQCSMA, the transmission schedule is determined according to the channel states. Let be the transmission schedule in the most recent data slot (before and including time ) when . More specifically,
is the transmission activity of the nodes in the last instance of time, before and including , where the channel state vector realization was . Consider the following example to better illustrate the evolution of . Assume that only the channel of node is time varying. Since there is only one time varying channel, there are only two variables defined by , i.e., and . The sample path evolution of , and are given in Figure 5. Suppose that at time , RQCSMA has chosen to schedule an idle slot. Note that since at , the channel state is , updates it value to the value of while maintains its value from , since it is not allowed to change its value. Basically, when , only evolves, and does not evolve. Note that in scheduling the nodes, if (), RQCSMA uses ().
At time slot , based on the realization of the channel state vector a decision schedule, is randomly determined. Then each node searches for the corresponding to the channel state vector , i.e., . If node , and for all , then node chooses to be active with probability , and idle with probability . If for any , then will be inactive in the current transmission slot. Any node not selected in will maintain its previous state, i.e., it will be inactive if and active otherwise. The conditions on activation probabilities will be specified in Section 4.3. Note that in QCSMA, transmission schedule is determined based on the previous transmission schedule , whereas in RQCSMA we use for each to determine the transmission schedule . In Section 4.3, We will show that defining as such, and using it to schedule the transmissions, results in Markov chains with productform stationary distributions. This is then used to prove throughput optimality of RQCSMA. The detailed description of RQCSMA is given in Algorithm ?.
We illustrate the operation of RQCSMA with the following simple example.
Example 1
Suppose that the network consists of a source and a relay node with only node having a time varying channel, i.e., . Note that whenever , the network consists of two nodes willing to transmit to a common destination as in Figure 3. In this scenario, the possible decision schedules are , which translates into two nodes being idle, or direct transmission of node to destination or transmission of node to destination, respectively. Meanwhile, if , the network consists of a node which forwards its packet to the node for eventual delivery to the destination as shown in Figure 4. In this scenario, the possible decision schedules are , which translates into two nodes being idle, or node relaying a packet to node or transmission of node to destination, respectively.
Since there is only one time varying channel, we need to define , the transmission schedule corresponding to the last instance that and , the transmission schedule corresponding to the last instance such that according to (Equation 3). Note that is associated with Figure 3 and is associated with Figure 4.
First consider the case when , , i.e., in the most recent time slot, , when , where , both nodes did not transmit. Then,
if the decision schedule in the current slot is , then node transmits to the destination with probability and remains idle with probability .
if the decision schedule is , then node transmits to the destination with probability , and it remains idle with probability .
Now, consider the other possibility for the channel state of node , i.e., when , we have , which means that in the most recent time slot, , when , node was idle whereas node transmitted. Then,
if the decision schedule in the current slot is , then node does not transmit, because the last time when the node to destination channel was OFF, node was active.
if the decision schedule in the current slot is , then node transmits to the destination with probability , and it remains idle with probability .
3.3Contention Resolution Scheme
In this section, we propose a contention resolution scheme which satisfies two purposes: i) Randomly choosing a feasible schedule, and ii) Inferring the channel state vector at each node in a distributed manner. At the beginning of contention slot, the destination node broadcasts a probe packet. Upon successful detection of this packet, node determines that its channel to the destination node is ON at slot , and otherwise, it recognizes that its channel is OFF. We assume that the contention slot is divided into number of mini slots. Each node is registered to one of the first mini slots, i.e., we assign , , mini slot to node. At mini slot, node broadcasts a random signal, and all other nodes attempt to detect the random message. If they detect a message on minislot, they conclude that , otherwise . Hence, at the end of minislots, all nodes are aware of other nodes’ channel states to the destination.
To randomly choose (i.e., with probability for any ) a decision schedule , node (given that ) randomly selects a number uniformly distributed in , and waits for mini slots. If node hears an INTENT message before , it will not be included in . Otherwise, node broadcasts an INTENT message at the beginning of the control minislot. If there is no collision in control minislot, node will be included in . Finally, if there is a collision, none of the nodes will be included in (i.e., ).
Note that the contention resolution scheme requires all nodes to hear each other which may not be always feasible. This is similar to wellknown hidden node problem in wireless networks which can be resolved by RequestToDecide (RTD) and ClearToDecide (CTD) mechanism. The INTENT message in our original scheme is split into RTD and CTD pair transmitted and received, respectively, in two successive submini slots. In the first submini slot, node , sends RTD to the destination node. If the destination receives RTD without a collision (i.e., no other nodes are transmitting in the same submini slot), then the destination replies with CTD to node in the second submini slot. If node receives the CTD from destination without a collision, then node is added to the decision schedule.
4Throughput Optimality of RQCSMA
In this section, we first discuss why a direct application of QCSMA is not possible for our relay network. Then, we explain how we can modify the standard analysis of throughput optimality according to the structure of our proposed RQCSMA algorithm. Finally, we provide the proof of throughput optimality of RQCSMA when the channel states are determined correctly according to the contention algorithm discussed earlier. The key notations used in this paper are listed in Table 1.
Notation  Description  

Channel state between node and the destination at time ,  
Number of packets arriving to node at time ,  
Size of the queue, storing the packets of node at time ,  
Relaying queue of node at time ,  
Scheduled node at time ,  
The decision schedule,  
The decision schedule when ,  


Activation probability of node when ,  
Weight associated with scheduled node at time .  
4.1Drawback of QCSMA
Even though QCSMA is known to achieve a throughputoptimal schedule in a distributed fashion for a wireless network with nontimevarying links, it is not a suitable choice for many other network models. A particular case was studied in [49], where the authors investigated a random access scheme for scheduling nodes in a CRN. The authors in [49] demonstrated that QCSMA is not throughput optimal in this network model, since the Discrete Time Markov Chain (DTMC) with the transmission schedule chosen as the state is not timereversible, and thus, it does not have a productform solution.
Similarly, QCSMA is not a suitable choice for our system model as well. Note that QCSMA is throughput optimal when the activation probabilities for each node are chosen to be a function of the node weights as defined in (Equation 2) and ( ?) of [29]. It can be argued that the stationary distribution of DTMC exists in steady state, if the activation probabilities change slowly over time. Unfortunately, in our system model, the node weights may vary abruptly depending on the channel state. Specifically, in Example 1, whenever , the weight of node is and it becomes otherwise. Consider the case where and are both very large and equal. In this case, the activation probability of node is approximately equal to when and otherwise. This results in a transition probability jumping from to from one slot to the next. This implies that the distribution of DTMC cannot be assumed to be at steady state at a given time
4.2Modifying the DTMC with channel state information
Next, we demonstrate that associating the channel state of the node into the states cannot resolve this problem. For this purpose, we again consider the network in Example 1. Let be the states of the new DTMC as shown in Figure 6
Inspired by the approach of [49], we define a separate DTMC for each of the different realizations of channel state vector to mitigate this problem. Let be the states of DTMC. Note that a transition from state to another state in DTMC only depends on the current state given the current channel state vector and the decision schedule. Thus, DTMC is Markovian for all .
Figure ? shows two different DTMCs associated with each of two different states of for the network in Example 1. The transition probabilities follow directly from the definition of RQCSMA algorithm. Figure 7 represents DTMC where a “Rectangle” corresponds to and an oval corresponds to . For example, consider in DTMC; being in this state means that the channel of the node is OFF and only node is transmitting, while means that the channel of the node is ON and the last time that channel was OFF prior to the current time slot, node was transmitting. This follows from the definition of in (Equation 3). Similarly, Figure 8 represents DTMC where a “Rectangle” corresponds to and an oval corresponds to . Note that according to RQCSMA, any state, , (i.e., a rectangle) in DTMC at any time can make a transition into another state, (), if and only if , and the state freezes otherwise. The same argument follows for DTMC.
Now consider and as states for DTMC and DTMC, respectively
4.3Proving the Throughput Optimality of RQCSMA
In the following, we consider the general network topology with a source and relay nodes, with all channels to the destination being time varying ONOFF channels, whereas the channels from node to all other nodes are always ON. We show that DTMC for all is time reversible with a productform stationary distribution. Theorem ? gives the productform of the stationary distribution, and finally, Theorem ? claims the throughput optimality of RQCSMA.
Consider the state transitions of DTMC as given in Figure 9. For any we can write the following detailed balance equation which follows from RQCSMA:
which simplifies to:
Next we have,
If we define , we have:
Substituting (Equation 5) in (Equation 4) gives the stationary distribution of DTMC.
The productform stationary distribution of DTMC, suggests that we can use the results established in [13] to prove the throughput optimality of the algorithm. Let be the schedule at time slot . For the sake of simplicity, we drop and denote the weight associated with the scheduled node by . Similarly, is the weight associated with the .
[th1]Theorem
If we choose , then the stationary distribution of DTMC becomes:
By choosing ’s in and wisely, changes slowly over time and we can assume that DTMC is in steadystate in every time slot (time scale separation) [29]. In the following we show that RQCSMA is throughput optimal by showing that it is close enough to another throughput optimal algorithm (i.e., MWS). [th1]Theorem
We define . Next define the following states:
Note that so . We now calculate the probability of a state in set :
where (Equation 6) is true because , and
Therefore, if
then . Then we have the following results:
Let be the node scheduled in slot by RQCSMA. Using the total probability law we have:
where . Note that holds because and whenever . Hence, Algorithm ? satisfies the condition of Theorem ?, and thus, it is throughput optimal.
4.4On the complexity of the RQCSMA
At time slot , each node first needs to to find a decision schedule, and the channel state vector . Note that this requires mini slots which results in complexity. Then in RQCSMA each node needs to find the appropriate DTMC scheduled for evolution which means that each node needs to find a single element from a set with size . It is well known that the simple binary search, has logarithmic complexity [51]. Consequently, the total complexity of RQCSMA is .
5Achievable Rate Region of the Relay Network
In order to validate the throughput optimality of RQCSMA, we focus on a network consisting of a source and a relay nodes. This helps us visualize the geometric representation of the achievable rate region and simplify the performance comparison of RQCSMA and QCSMA. Lemma ? establishes the achievable rate region for two nodes with node having arrival rate of and probability of ON channel state, and node having arrival rate of and probability of ON channel state, :
The proof is given in Appendix.
The achievable rate region of the relay network, , for the case of two nodes with and with is depicted in Figure 10 and Figure 11, respectively. Note that in case of no relay service, node can only achieve a rate of , whereas when relaying is enabled, it can achieve an additional rate of for and for .
6Numerical Results
In this section, we numerically validate the throughput optimality of RQCSMA and compare its achievable rate region with those of QCSMA and a simple algorithm utilizing uniform random backoff [29]. The uniform backoff (UB) algorithm is defined in Algorithm ?. Note that QCSMA is not designed for time varying channels so in QCSMA if a node is chosen to be activated when its channel is at OFF state, it remains idle in that slot. Also, note that we choose ’s in (Equation 2) and ( ?) to be of the form with .
Note that we do not employ the RTD/CTD mechanism in the simulations, since it is required to be implemented by other policies as well to solve the hidden node problem, and its overhead will be similar in all. Hence, in the following we assume that any pairs of nodes are able to hear each other.
6.1Validating Throughput Optimality by Simulations
We consider a network consisting of two nodes. We evaluate the performance of RQCSMA according to the average queue lengths of all queues in the network, i.e., . To calculate , we run the simulation for slots, and over 10 different sample paths. The channel statistics are chosen as , . The arrival rates, , are taken in the region where we evaluate for each point in this region. Figure 12 shows the contour graph of versus and , where the solid and thick black line is the achievable rate region of the relay network. We observe that the network under RQCSMA exhibits unstable behavior, shown by the increase in the average queue lengths (lighter colored contours) as we cross the boundary of the stability region.
Next, to further investigate the stability region of the RQCSMA, we plot the time evolution of the queues at certain points in the achievable rate region. We choose points , and on the boundary of the achievable rate region as depicted in Figure 12. Then, we evaluate the time evolution of the queues at , and , i.e., the points are inside the achievable rate region and they are sufficiently close to the boundary where . Also we evaluate the time evolution of the queues for , and , i.e., the points are outside the achievable rate region and they are sufficiently close to the boundary. Figures Figure 13, Figure 15 and Figure 17 show that whenever we take a point inside the achievable rate region close to the boundary, RQCSMA is able to stabilize the queues as we expect from the throughput optimality of RQCSMA.
We repeat the same procedure for QCSMA and UB algorithms. Figure 19 and Figure 21 show the contour graphs of versus and under QCSMA and UB, respectively. We observe that the network under both QCSMA and UB exhibit unstable behavior, shown by the increase in the average queue lengths, even when the arrivals are far away from the boundaries of the stability region. To investigate further, we plot the time evolution of the queues under QCSMA and UB at . The results are depicted in Figure 20 and Figure 22. It can be seen that QCSMA and UB are unable to stabilize the queues for arrival rates stabilized by RQCSMA.
6.2The effect of the number of relay nodes
Next, we numerically compare the performance of RQCSMA with UB, MWS and QCSMA algorithms when there is more than one relay node. For MWS algorithm, at each time slot, we choose an action for each node which maximizes the weights as defined in (Equation 2) and ( ?). We evaluated the average queue lengths of all nodes for a given arrival rate vector as our performance measure and illustrate confidence intervals.
We consider a network with a source and three relay nodes. We choose , , and for the channel state statistics.
We compare the performance of the algorithms in terms of average queue lengths of all nodes. We begin with an arrival rate vector and linearly increase arrival rate of node from to . For each algorithm, for a fixed we run 10 independent simulations and take the average. The average queue lengths of nodes with respect to is shown in Figure 23, where the vertical bars represent the confidence interval. It can be seen that RQCSMA outperforms UB which does not take into account the queue length information and also QCSMA which does not take channel state information into account.
Although the average queue length information gives us an idea about the performance of these algorithms, it does not say much about the stability of the queues. Hence in the following, we compare the algorithms in terms of their queue length evolution with respect to time for a given arrival rate vector.
It can be seen from Figure 23 that for (i.e., ) there is an increase in the average queue lengths of all nodes in all algorithms suggesting that the network is operating at a point close to the boundary of its achievable rate region. We pick the arrival rate vector as where the increase in average queue lengths of nodes is apparent in MWS algorithm according to Figure 23, i.e., is close to the boundary of the stability region. We again run each algorithm for arrival rate vector for 10 random seeds and then we take their average. The evolution of the queue lengths of nodes (where the sizes of all queues are summed) versus time is given in Figure 24, where the vertical bars represent the confidence interval. It can be seen that for the desired arrival rate vector, MWS and RQCSMA can stabilize the queues whereas in UB and QCSMA the queue sizes tend to increase as the time evolves suggesting the instability of the queues.
7Conclusion
In this paper, we addressed the problem of scheduling in wireless relay networks with a source and multiple relay nodes where all nodes transmit to a common destination node over independent ONOFF channels. In this relay network, relay nodes help a single node with bad channel quality by relaying its packets. The scheduling is based on a variant of well known QCSMA algorithm, and we prove the throughput optimality of the developed algorithm. We proposed a new contention resolution algorithm, which is used by the nodes to infer which nodes have ON channels; an information required by the scheduling algorithm for achieving throughput optimality. RQCSMA is different from QCSMA in the sense that it uses channel state information to schedule the nodes and it actually benefits from time variability of the channel states. The performance of the RQCSMA, QCSMA, MWS and a simple random backoff based algorithms are compared through simulations to demonstrate the efficacy of RQCSMA.
As a future work, we will consider the extension of RQCSMA to multihop and multi source nodes relay networks. Also note that the throughput optimality of RQCSMA depends on the assumption that nodes are able to perfectly infer the channel state vector in the contention slot. The investigation of scheduling algorithms with partial and incomplete channel state information is left as future work.
Proof of Lemma
To simplify the proof, we assume that node has a single queue and stores the received packets from node in its own queue. Note that this does not change the stability region of the network. Define and as the instantaneous service rates of node and node , respectively, and as the instantaneous service rate of node if it relays its packets to node . Note that , and take binary values or and satisfy the following inequality . Then, the evolution of the queues are given as:
It is well known that when MWS is applied the resulting service rates are throughput optimal. Thus, the expected service rates can be calculated as follow:
Denote . The drift in the Lyapunov function, , can be written as:
Now define as a bound on the summation of first and third terms above as follows:
Now let us assume that so that the achievable rate region is as in ( ?). First we focus on and . For small positive value of we have . We write the conditional Lyapunov drift as follow:
With a similar approach we can show that: