Resource allocation based on hybrid genetic algorithm and particle swarm optimization for D2D multicast communications
Graphical abstract
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 . cellular user equipments (CUEs) , and D2D pairs , are uniformly distributed within that circle. We denote 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 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 . 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)
- et al.
An efficient genetic algorithm for large-scale transmit power control of dense and robust wireless networks in harsh industrial environments
Appl. Soft Comput.
(2018) - et al.
Proposing a centralized algorithm to minimize message broadcasting energy in wireless sensor networks using directional antennas
Appl. Soft Comput.
(2018) - et al.
Hybrid multi-objective evolutionary algorithms based on decomposition for wireless sensor network coverage optimization
Appl. Soft Comput.
(2018) - et al.
Salp swarm algorithm: A bio-inspired optimizer for engineering design problems
Adv. Eng. Softw.
(2017) - et al.
Grey wolf optimizer
Adv. Eng. Softw.
(2014) - et al.
Binary dragonfly optimization for feature selection using time-varying transfer functions
Knowl.-Based Syst.
(2018) - et al.
Joint channel allocation and power control based on pso for cellular networks with d2d communications
Comput. Netw.
(2018) - et al.
Allocation optimization based on multi-population genetic algorithm for d2d communications in multi-services scenario
- et al.
An intelligent system for spam detection and identification of the most relevant features based on evolutionary random weight networks
Inf. Fusion
(2019) - et al.
Binary dragonfly optimization for feature selection using time-varying transfer functions
Knowl.-Based Syst.
(2018)