Elsevier

Applied Soft Computing

Volume 83, October 2019, 105605
Applied Soft Computing

Resource allocation based on hybrid genetic algorithm and particle swarm optimization for D2D multicast communications

https://doi.org/10.1016/j.asoc.2019.105605Get rights and content

Abstract

Device-to-Device (D2D) multicast communication is a new technology of the fifth-generation networks (5G) for efficiently coping with the ever-increasing demand for content sharing among users. In fact, it enables direct communication between devices in proximity and improves spectral efficiency by reusing a licensed cellular spectrum. The existing related studies show that D2D communications increase network capacity and reduce latency. Nevertheless, the interference management should be carried out in a coordinated manner in order to realize the full potential of this technology and enable its integration into the cellular architecture. We consider the joint uplink subcarrier allocation and power control in D2D underlying cellular networks. In single rate multicast communications, the achieved data rate is greatly limited by the nodes with low channel quality. In this article, we formulate the resource allocation as a max–min optimization problem. Such an optimization problem is in general an NP-hard combinatorial problem and its solution typically requires searching enormous search trees. We propose a multicast schemes based on Genetic Algorithm (GA) and Binary Particle Swarm Optimization (BPSO). We implemented 8 different transfer functions combined with 2 update position strategies in order to assess the performance of BPSO. For the GA, we implemented Uniform and Multipoint crossover methods combined with Roulette Wheel and Tournament selectors. In addition to providing a deep understanding of the algorithm behavior, we present numerical results that demonstrates that GA outperforms BPSO in terms of minimum achieved data rate when the pressure of infeasibility is high.

Introduction

With the increasing development of multimedia services and the expected explosion in the number of wireless devices, it is a great challenge for the existing cellular network to satisfy the users’ demand [1]. Most of group communication applications such as video streaming and multiplayer gaming require high data rate and low latency. Hence, multicasting is an important network feature that is required by these applications. In single-rate multicast communications, base stations broadcast the same information to a group of users at a common rate. The links experiencing poor channel conditions lead to throughput bottleneck and the users are unable to correctly receive the relevant source information. Thus, the differentiation of channels quality results in bandwidth waste for users with better channel quality [2]. The users with the same demand and in proximity to each other may cooperate in order to improve the data dissemination efficiency. The devices with high quality link are used as relays for multicast retransmissions and low-rate links are bypassed by substituting them with high-rate two-step multicast.

Device-to-Device (D2D) communications exploit the proximity of mobile devices in order to exchange information over direct links without the need for routing via a base station. D2D communications underlaying cellular network hold the promise to improve both spectral utilization efficiency by sharing the spectrum of cellular users and low transmission power by transmitting over short range links . Introducing D2D multicast communications to conventional cellular networks can not only alleviate the spectrum scarcity problem but also offload the base station. D2D users, acting as relay nodes, participate in lowering the burden of base stations by retransmitting data with high data rate through direct and short-range links.

Even though the D2D users can transmit and receive signals directly, they remain under the control of the base station. Sharing spectrum between the cellular users and the D2D users induces severe mutual interference. Interference management is one of the most critical issues for D2D communication enabled in cellular networks. Without effective interference mitigation, resulting in communication interruptions, the benefits brought by D2D communications would be eliminated [3]. In the literature, there has been numerous studies to investigate resource management in D2D multicast communications with different objectives and constraints [4], [5], [6], [7]. Different from the existing works in [8], [9] and [10], this paper investigates a general case in which joint optimization of subcarrier allocation and power control for D2D transmitters are considered. The main target is to optimize the minimal achieved data rate by DUEs while guaranteeing the basic Quality Of Service (QoS) requirements of CUEs. In our system model, a DUE is allowed to use multiple subcarriers and a subcarrier can be allocated to multiple DUEs. We formulate the problem as a mixed integer nonlinear programming (MINLP) problem. It combines the challenges of handling the non-linearities with the combinatorial explosion of integer variables. This renders the algorithm design a tough work. We propose a resource allocation scheme based on metaphors of swarm intelligence and evolutionary computation. Bio-inspired algorithms have been proven effective to solve diverse problems in wireless networks such as transmit power control of dense wireless networks in industrial environments [11], broadcasting energy optimization in wireless sensor networks [12] and coverage control optimization for wireless sensor network [13]. One of the most important class of Evolutionary algorithms is Genetic algorithm (GA). The search starts from a randomly generated population that evolves over successive iterations. GA uses various biological operations such as selection, crossover, mutation and reproduction in order to propagate its population from one generation to another.

Swarm Intelligence (SI) techniques mimic the swarming behavior of organisms that live in groups and cooperate among themselves. Examples of SI algorithms include but not limited to Particle Swarm Optimization (PSO) [14], Ant Colony Optimization (ACO) [15], Dragonfly Algorithm (DA) [16], Salp Swarm Algorithm (SSA) [17] and Grey Wolf Optimizer (GWO) [18]. The PSO is the dominant SI algorithm that has been widely used in the literature. These algorithms were originally designed to solve problems with continuous variables. Transfer functions can be used to convert the continuous optimizer to suit binary problems where the search space is represented by binary values [19].

To sum up, the main contributions of this paper are summarized as follows.

  • 1.

    We consider a general system model, in which we have no restrictions on the number of subcarriers used by a D2D user and no assumptions on the number of D2D users using a specified subcarrier. The problem of joint subcarrier allocation and power control is, thus, formulated as a mixed integer nonlinear programming which maximizes the minimal data rate for D2D users imposing a minimum SINR for cellular users.

  • 2.

    Another major challenge is that the subcarrier allocation and power control are mutually coupled. To make the optimization problem tractable, we consider two separate decision variable vectors. A vector of binary variables represents the subcarrier allocation and a vector of continuous variables defines the power control. These two decision vectors are conducted by two algorithms.

  • 3.

    We propose a flexible resource allocation scheme relying on Particle Swarm Optimization and Genetic Algorithm. GA is used to solve the combinatorial issue of subcarrier allocation. The individuals in the population are evaluated using a feedback from the PSO algorithm which computes the near-optimal solution for the power control problem.

  • 4.

    We evaluate the hybrid GA–PSO scheme performance. It is demonstrated via simulation that using our new algorithm outperforms the existing approach based on standard PSO for both power control and subcarrier allocation for a particular position update policy and network settings characterizing high infeasibility rate.

The rest of this paper is structured as follows. Section 2 presents the related work and Section 3 introduces the system model and problem formulation. The proposed algorithm is described in Section 4. In Section 5, numerical results and performance analysis are provided. Finally, we conclude this investigation in Section 6.

Section snippets

Related work

An investigation of the current literature suggests that the existing works in D2D resource allocation have largely considered the unicast scenario [20], [21], [22], [23]. However, the multicast scenario has its own challenges. Having a user equipment instead of the base station performing multicast limits the capabilities of the network. Our review of the research works dedicated to resource allocation for underlying D2D multicast communications mainly focuses on the objective function

System model and problem formulation

The considered architecture includes two tiers. The first tier includes communication between the Base Station (BS) and the cellular users and the second tier involves D2D communication. We consider a single-cell environment where the BS is located in the center of a circle with a radius of R. M cellular user equipments (CUEs) m=1,,M, and N D2D pairs n=1,,N, are uniformly distributed within that circle. We denote r as the maximal distance between a D2D transmitters and its corresponding D2D

Genetic algorithm

The genetic algorithm (GA) is an iterative stochastic optimization method inspired by the natural evolution that uses biological metaphors to generate new points in the search space. A formal model for the effectiveness of the GA search process is defined in the GA Schema Theorem formalized by Holland [38] and popularized by Goldberg [39].

The GA encodes a potential solution on chromosome-like data structure. A set of potential solutions (population) is randomly initialized. Each individual is

Simulation results

In this section, we perform a series of simulations to evaluate the performance of the hybrid GA–PSO algorithm . As shown in Fig. 3, the simulations are carried out in a single cell where the D2D pairs and the cellular users are randomly distributed in a circle with radius 500 m. The main simulation parameters are listed in Table 1. We aim to analyze the effect of hybridization on the obtained results. Therefore, we compare our hybrid GA–PSO algorithm with PSO algorithm. In PSO algorithm, the

Conclusion and future directions

In this article, we proposed an hybrid GA–PSO for joint subcarrier allocation and power control in D2D multicast underlay network. The discrete allocation for subcarriers is performed via the GA and the continuous standard PSO is applied to the power control problem. A monotonic ascending iteration process converging in a finite number of rounds has been analyzed. The performance of the hybrid GA–PSO method is put in contrast the PSO algorithm where the discrete binary PSO is used to solve the

Acknowledgment

This research was funded by the Deanship of Scientific Research at Princess Nourah bint Abdulrahman University through the Fast-track Research Funding Program.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (59)

  • HeidariA.A. et al.

    An efficient modified grey wolf optimizer with Lévy flight for optimization tasks

    Appl. Soft Comput.

    (2017)
  • MafarjaM. et al.

    Binary grasshopper optimisation algorithm approaches for feature selection problems

    Expert Syst. Appl.

    (2019)
  • MafarjaM. et al.

    Evolutionary population dynamics and grasshopper optimization approaches for feature selection problems

    Knowl.-Based Syst.

    (2018)
  • FarisH. et al.

    An efficient binary salp swarm algorithm with crossover scheme for feature selection problems

    Knowl.-Based Syst.

    (2018)
  • TreleaI.C.

    The particle swarm optimization algorithm: convergence analysis and parameter selection

    Inform. Process. Lett.

    (2003)
  • MirjaliliS. et al.

    S-shaped versus v-shaped transfer functions for binary particle swarm optimization

    Swarm Evol. Comput.

    (2013)
  • DahiZ.A.E.M. et al.

    On the efficiency of the binary flower pollination algorithm: application on the antenna positioning problem

    Appl. Soft Comput.

    (2016)
  • ShahS.T. et al.

    Device-to-device communications: A contemporary survey

    Wirel. Pers. Commun.

    (2018)
  • HamdiM. et al.

    Ga-based scheme for fair joint channel allocation and power control for underlaying d2d multicast communications

  • ToumiS. et al.

    An adaptive q-learning approach to power control for d2d communications

  • MeshgiH. et al.

    Joint channel and power allocation in underlay multicast device-to-device communications

  • MeshgiH. et al.

    Optimal resource allocation in multicast device-to-device communications underlaying lte networks

    IEEE Trans. Veh. Technol.

    (2017)
  • KimJ.H. et al.

    Resource allocation for multiple device-to-device cluster multicast communications underlay cellular networks

    IEEE Commun. Lett.

    (2018)
  • WuX. et al.

    Joint resource allocation and power control for cellular and device-to-device multicast based on cognitive radio

    IET Commun.

    (2014)
  • SunS. et al.

    Device-to-device resource allocation in lte-advanced networks by hybrid particle swarm optimization and genetic algorithm

    Peer-to-Peer Netw. Appl.

    (2016)
  • LiangT. et al.

    Joint resource allocation and power control scheme for device-to-device communication underlaying cellular networks

  • SunS. et al.

    Resource allocation for d2d communication using particle swarm optimization in lte networks

  • KennedyJ.

    Particle swarm optimization

  • DorigoM. et al.

    Ant colony optimization: a new meta-heuristic

  • Cited by (0)

    View full text