Mejdi Kaddour

Assigning and scheduling partially overlapping channels in wireless mesh networks

The efficiency of multi-channel multi-radio wireless mesh networks can be improved with the increase of the number of channels and radios. Despite the availability of up to 11 channels in 802.11, we can choose only up to three non overlapping channels at any given time. In this study, we investigate how to design a channel assignment and a scheduling algorithm, which both exploit partially overlapping channels in order to maximize the throughput. We next examine how much additional throughout we can obtain by doing so, in comparison with only using three orthogonal channels. Numerical experiments show that we can gain up to 25% of throughput by appropriately managing overlapping channels.

A Column Generation Approach to Maximize Capacity of Multi-rate Power Controlled TDMA Wireless Sensor Networks

This paper describes a mixed-integer linear programming model to maximize wireless sensor networks (WSN) capacity by determining traffic routes and target coverage, adjusting data rates, and dynamically controlling transmission power. Radio transmissions are studied under signal-to-interference-plus-noise-ratio (SINR) model. In the considered WSN, a set of targets needs to be continuously monitored by a given number of already deployed sensors. The objective is to gather as much sensing data as possible during a unit time period. Since this type of problems is known to be NP-hard, we propose a computationally feasible column generation based approach to find near-optimal solutions when dealing with integer flow and coverage variables. Extensive computational experiments even show that optimal solutions are reached in most cases for reasonable network sizes. Numerical results illustrate the benefits from extending the number of available transmission power levels, the major impact of data rate adaptation on throughput, and demonstrate that achieving higher network capacity comes at the expense of deploying more sensors than the minimum required to connectivity and target monitoring.

An efficient method to minimize TDMA frame length in wireless sensor networks

We address the problem of minimizing TDMA frame length in a wireless sensor network charged with monitoring a given set of targets. The problem reduces to finding a minimal set of so-called configurations that delivers data from the targets to a specially designated sink node, where each configuration is a set of links that can transmit concurrently without significant interference. We assume a SINR-based interference model. We use a column generation technique to derive near-optimal solutions even when integrality constraints are enforced on coverage and flow variables. We also introduce a heuristic and a hybrid algorithm to efficiently solve the problem by leveraging a wide range of network parameters related to transmission power control, data rates, and routing. Our results show that significant gains in scalability can be obtained by using our methods.

Refining the impact of partially overlapping channels in wireless mesh networks through a cross-layer optimization model

Although many works exist on the subject, the real impact of partially-overlapping channels (POCs) on the capacity of multi-radio wireless mesh networks is not clearly established yet. In this paper, we propose a mathematical programming model to evaluate accurately the improvement in throughput that could be achieved if all the adjacent radio channels were exploited. Compared to previous works, our model tackles the problem with a fine-granularity control of a wider set of radio and network parameters related to routing, dynamic channel assignment, rate adaptation, transmission power and link scheduling. Due to its complexity, our problem is solved using column generation. Furthermore, we propose a suboptimal fast procedure to solve the induced subproblem of link selection with joint channel assignment, rate adaptation and power control. While our numerical experimentations on random topologies show that employing POCs leads to similar throughput increase (i.e., around 11%) to some previous work, our main result is that this gain is found to be almost negligible for more realistic topologies.

On TDMA scheduling in wireless sensor networks

We study the problem of finding an efficient schedule for convergecast in a wireless sensor network, using the SINR model of interference, and a TDMA protocol for medium access. We compare two approaches to computing a minimum length schedule for the TDMA frame. In the first approach, called the TC-approach, a multi-set of transmission configurations that are interference-free and that cover the convergecast traffic is computed and the scheduling algorithm restricts itself to using these configurations. In the second approach, called the tree-based approach, first a routing tree or subgraph is computed, and next, sets of non-interfering links are scheduled in rounds, based on which links have available traffic in each round. In this paper, For the TC-based approach, we provide new column generation approach and several new scheduling heuristics. For the tree-based approach, we propose the construction of a new tree called TFM tree, which takes into account variable power assignment, as well as a new scheduling algorithm for the second phase, called the ROS algorithm. We performed extensive experimental evaluations of both approaches. Our results show that the tree-based approach using the TFM tree significantly outperforms the TC-approach.

Increased throughput and reduced delay with partially overlapping channel in WMNs

We instigate the questions of: (i) whether increasing the number of available channels is always useful for improving the Quality of Service (QoS) (i.e., throughput or packet delivering delay) of Wireless Mesh Networks (WMNs), (ii) whether using more complex algorithms for managing Partially Overlapping Channels (POCs) is beneficial in terms of improving the QoS of WMNs? For the purpose of the instigation, we design a set of algorithms that can be combined to increase the spatial reuse and to schedule efficiently packet transmissions in WMNs. Extensive numerical experiments indicate that using POCs leads not only to a significant increase in network throughput as reported in previous studies, but also to a considerable decrease in average delay.

Dimensioning microwave wireless networks

We aim at dimensioning fixed broadband microwave wireless networks under unreliable channel conditions. As the transport capacity of microwave links is prone to variations due to, e.g., weather conditions, such a dimensioning requires special attention. It can be formulated as the determination of the minimum cost bandwidth assignment of the links in the network for which traffic requirements can be met with high probability, while taking into account that transport link capacities vary depending on channel conditions. The proposed optimization model represents a major step forward since we consider dynamic routing. Experimental results show that the resulting solutions can save up to 45% of the bandwidth cost compared to the case where a bandwidth over-provisioning policy is uniformly applied to all links in the network planning. Comparisons with previous work also show that we can solve much larger instances in significantly shorter computing times, with a comparable level of reliability.

A Multicast Routing Protocol adapted to the characteristics of Wireless Mesh Networks

Wireless Mesh Networks are a wireless multi-hop networks in which the nodes are characterized by theirs stability. In this sort of network the minimum cost tree connects sources and receivers by implying a minimum number of forwarding nodes. This paper introduces an adaptation to the popular multicast protocol ODMRP (On-Demand Multicast Routing Protocol) named OODMRP (Optimized On-Demand Multicast Routing Protocol). The main effect of OODMRP is to minimize the number of forwarding nodes which leads to the optimization of resources consumption and the network congestion. Indeed, before selecting a forwarding node, each node looks in its neighborhood for the existence of a forwarding node. In this case, it selects this node as an ascending one in the multicast tree. Our simulation results show that OODMRP reduces dramatically the number of forwarding nodes and forwarded packets, compared to original ODMRP.

Cost-effective Bandwidth Provisioning in Microwave Wireless Networks under Unreliable Channel Conditions

Cost-effective planning and dimensioning of backhaul microwave networks under unreliable channel conditions remains a relatively under explored area in the literature. In particular, bandwidth assignment requires special attention as the transport capacity of microwave links is prone to variations due to, e.g., weather conditions. In this paper, we formulate an optimization model that determines the minimum cost bandwidth assignment of the links in the network for which traffic requirements can be fulfilled with high probability. This model also aims to increase network reliability by adjusting dynamically traffic routes in response to variations of link capacities induced by channel conditions. Experimental results show that 45% of the bandwidth cost can be saved compared to the case where a bandwidth over-provisioning policy is uniformly applied to all links in the network planning. Comparisons with previous work also show that our solution approach, based on column generation technique, is able to solve much larger instances in significantly shorter computing times (i.e., few minutes for medium-size networks, and up to 2 hours for very large networks, unsolved so far by previous models/algorithms), with a comparable level of reliability.

Exact and Efficient Heuristic Deployment in WSN under Coverage, Connectivity, and Lifetime Constraints

Wireless sensor networks lay down many challenging optimization problems, such as coverage, nodedeployment,trackingorenergyconservation.Inthispaper,weareinterestedindeployment strategiesthatresultinaminimumofsensorsnetworkwhileensuringtargetcoverageconnectivity betweenthesensorsandsink.Tothisend,weproposetwoalternativedeploymentapproachesbased onintegerlinearprogrammingandweexploitthelinear-programmingsequentialfixingtechnique todesignthreepolynomial-timeheuristicprocedures.Theperformanceandeffectivenessofthese approachesintermsofnetworkcostandcomputationalrequirementsarehighlightedthroughseveral experiments.Furthermore,weinvestigatethenetworklifetimeproblemwhereagivenoperational durationmustbereached. KEyWoRDS Connectivity, Coverage, Energy, Integer Linear Programming, Lifetime, Wireless Sensor Network