Finding the Maximal Independent Set (MIS) in a graph is a well-known problem with applications in resource allocation, load balancing, and routing optimization. This task is particularly challenging for large graphs as it requires multiple iterations over the entire set of vertices. Recently, there has been significant interest in developing techniques to maintain the MIS dynamically in evolving graphs rather than re-computing from scratch. In this paper, we propose new data structures and techniques for computing MIS in parallel on dynamic graphs. We specifically propose techniques to handle insertions and deletions in a batched setting. We conducted detailed experiments on shared memory multicore CPUs using graphs ranging from 50 million to ${1. 2}$ billion edges. Our results show that using our technique for insertions and deletions can provide up to 15.64x and 10.57x speedups on average over comparable baselines. Additionally, the final MIS we produce varies by only about ${0. 1 8 \%}$ in cardinality compared to the existing state-of-the-art.

Device-to-device (D2D) communication is an emerging 5G technology to improve spectrum efficiency and reduce data traffic load in wireless networks. In this paper, a joint uplink-downlink resource allocation (JUDRA) approach is presented to maximize the network throughput while guaranteeing the quality of service (QoS) for both cellular users (CUs) and D2D users (DUs). In particular, the proposed JUDRA approach allows DUs to reuse either uplink or downlink CUs’ resources. Besides, an optimization problem is formulated as a mixed-integer nonlinear programming problem (MINLP) which is generally NP-hard. The problem is solved in two stages: sub-channel (SC) assignment stage and power allocation stage by utilizing the characteristics of sequential geometric programming. Moreover, a low computational complexity approach is suggested that converges to the near-optimal solution in polynomial time. Finally, comparative results are presented to validate the effectiveness of the proposed approach.

In this paper, we present a distance and interference based decision-making framework for device-to-device (D2D) communication mode selection (choosing one mode among dedicated mode, reuse mode, and cellular mode) and a power control approach for multiple D2D pairs underlaying cellular network. The proposed approach is based on the location of D2D users. If orthogonal resources are available and the D2D pair is close to each other, then the D2D dedicated mode is preferred. Otherwise, reuse (uplink or downlink) mode is utilized where interference is a prime concern. Reuse mode of D2D communication with allowable interference allows more users to be served with efficient spectrum utilization. Moreover, the power control approach is introduced to avoid the wastage of power and to reduce the impact of interference. Numerical results are provided to validate the significance of the proposed approach.

Small cell and heterogeneous network are the promising technologies to provide good coverage and throughput and hence, recently attracted attention for the future cellular network. However, backhauling of irregularly located large number of small cells can be a critical issue in its deployment. The paper provides a general framework of massive MIMO backhauling in a heterogeneous network. The architecture consists of full duplex small cells which can receive backhaul data from massive MIMO enabled macro base station and transmit data to its users in the same frequency band. The paper aims at jointly optimizing the sum logarithm rate with respect to user association, number of antennas, and bandwidth allocation, subject to wireless backhaul as a constraint. A mixed integer non-linear programming problem is formulated. To solve the problem efficiently, the primal problem is splitted into three sub-problems and solved individually. Simulation results demonstrate considerable improvement in the performance over the existing results.

This study addresses the energy efficiency (EE) maximisation problem in multi‐input multi‐output cognitive relay networks, where a secondary transmitter–receiver pair communicates through an amplify‐and‐forward relay using the licenced spectrum of a primary user. The authors aim to maximise the EE of the secondary system by jointly designing the source and relay precoding matrices, subject to transmit power, interference, and quality‐of‐service constraints. The resultant problem is non‐convex in the precoding matrices; it is converted into a much simpler vector‐valued problem using singular value decomposition. A combination of iterative optimisation, fractional programming, and Lagrangian multiplier is used to further tackle the problem. Numerical results are presented to evaluate the EE performance of the proposed scheme. Impact of transmit power, interference threshold, and minimum rate requirement on the EE is analysed.

Wireless mesh networks (WMNs) are emerging as a cost effective solution for last-mile broadband access. Multiple antennas or multiple input multiple output (MIMO)-based systems have the potential for improving the capacity of these networks manifolds. However, unique physical layer characteristics associated with MIMO-based systems require joint design among physical layer and upper layers. In this paper, we investigate the problem of jointly optimizing routing and stream control scheduling in a MIMO-based WMN, where multiple antennas are exploited to increase the capacity via spatial reuse (SR) and/or spatial multiplexing (SM). A column generation (CG) method is used to efficiently solved the problem. For the difficult pricing problem that arises in the CG procedure, we propose a computationally efficient heuristic algorithm. A combination of dual decomposition and gradient-based methods are used to decouple the network layer and link layer operations. Simulation results are presented to evaluate the performance of the proposed scheme.

Wireless mesh networks (WMNs) are increasingly being deployed as backbone wireless networks for broadband Internet access. Use of multiple radios and multiple channels (MRMC) can bring substantial throughput improvement to these networks. In addition, multi-antennas or Multiple Input Multiple Output (MIMO) technology has great potential for improving the throughput of WMNs via spatial reuse and/or spatial multiplexing. In this paper, we investigate the cross layer design of routing, scheduling, channel allocation, and stream allocation in MRMC capable WMNs, where nodes are equipped with multiple antennas. In this paper, we model the radio and interference constraints on MRMC capable MIMO-based WMNs. We then formulate a joint routing, scheduling, channel allocation, and stream allocation (JRSCSA) problem with the objective of satisfying the traffic demands of a set of end-to-end sessions such that network schedule time is minimized. A column generation-based method is proposed to solve the JRSCSA problem efficiently. Simulation results are presented to evaluate the network performance and it is shown that MRMC capable MIMO-based WMNs have significantly lower network schedule time compared to single-radio single-channel networks.

Wireless Mesh Networks (WMNs) have emerged as a cost effective and promising technology to provide broadband connectivity in several environments. In this work, we consider a cross layer optimal design of routing, medium access control (MAC) scheduling, and physical layer resource allocation for WMNs, where nodes are equipped with Multiple Input Multiple Output (MIMO) antennas. MIMO antennas can greatly improves the throughput by transmitting multiple independent data streams and/or suppressing the interference from neighboring links. To address the MAC and stream control at the physical layer, we consider the notion of Transmission Configurations (TCs). By incorporating the TCs in the cross-layer design, we formulate a Joint Routing and Stream Control Scheduling (JRSCS) problem as a linear program (LP), with the objective of finding a minimum length schedule that satisfies the traffic demand of a set of end-to-end sessions. A column generation based method is proposed to solve the JRSCS problem. The simulation results are presented to verify the performance of the scheme. We have also evaluated the impact of the physical layer capabilities such as number of antennas, single and multi-user radios on the scheduling time of the network.

In this paper, we propose a near optimum adaptive iterative equalizer receiver for interference suppression in convolutionally coded wideband CDMA (W-CDMA) systems. In the proposed near-far resistant receiver, conventional matched filter front end is replaced by adaptive minimum mean-square error (MMSE) filter, for the generation of sufficient statistics. The receiver employs an adaptive soft-input soft-output (SISO) multiuser detector and a bank of single user SISO channel decoders. In the first iteration, multiuser detector processes received samples from chip-matched filter (CMF) and provides soft bit estimates. After deinterleaving, these bit estimates are decoded by SISO channel decoders. We have used optimum maximum a posteriori (MAP) decoders. The soft output decoders provide soft outputs for both decoded and coded bits. The soft estimates of coded bits are then fed to a symbol spaced adaptive MMSE equalizer in second or higher iterations, which further provides soft interference cancellation. Using simulations, it is shown that through an iterative process the effects of multiple-access interference (MAI) and inter-symbol interference (ISI) can almost be completely overcome and near optimum performance can be obtained

In this paper we propose the use of affine projection adaptive (APA) filtering algorithm for multiuser adaptive minimum mean square error (MMSE) receivers for asynchronous wideband code division multiple access systems (W-CDMA) in multipath fading channels. Performance of APA algorithm is compared with least mean square (LMS), normalized least mean square (NLMS), and recursive least square (RLS) algorithms. It is found that APA algorithm outperforms LMS and NLMS. Though RLS, algorithm offers better convergence characteristics but at much higher computational complexity. Three adaptive multiuser receiver architectures are considered, all employ chip-matched filter (CMF) at their front end. In the first two architectures perfect estimation of time delay of individual user is assumed and one CMF is employed per user. The difference between two architectures is that, in the first the samples collected from CMFs are cross-coupled and in the second architecture samples collected for individual users are not combined. In the third architecture, there is no need to estimate the time delay, but fractional sampling is used. It is shown that first architecture outperforms the other two architectures.