Spectrum allocation by wave based adaptive differential evolution algorithm

Abstract

Resource allocation is very important in mobile communication. This paper studies the assignment of spectrum in cellular networks operated on orthogonal frequency division multiple access (OFDMA) system. Both network utility and fairness among all linked users are taken to measure quality of service in cellular networks. The problem is expressed as a maximization optimization model. Wave based adaptive differential evolution (WADE) algorithm is proposed to for spectrum allocation. The WADE algorithm adapts algorithmic parameters in wave propagation manner to accelerate convergence process of the algorithm. Simulation results show that WADE is more efficient than other compared algorithms. Moreover, it is effective to allocate spectrum resources in OFDMA-based cellular networks.

Introduction

With the rapid development and wide application of information technology, the fifth generation (5G) wireless communication network has put forward higher requirements for the efficiency, reliability and security of the system [1], [2], [3]. Cooperative communication technology can obtain cooperative diversity gain through the cooperative processing of each node in communication and resource sharing, which can effectively improve the performance of wireless communication network, and become a research edge in the field of wireless communication [4], [5]. Resource allocation is an important technology to improve the performance of cooperative communication system. It can allocate the resources of time, frequency, power, relay station and so on in the system. The physical layer security technology can make full use of its inherent broadcasting and fading characteristics. It can bring additional security from the perspective of information theory which makes it an important complement to the existing security mechanism. Resource allocation is needed in sensor networks, ad hoc networks, cellular networks, etc. [6], [7], [8], [9].

OFDMA technology has high efficiency in spectrum utilization, and can be an effective method of solving multi-path fading [10]. It is an important multiple access technology in current and 5G communication networks. Robust resource allocation problems in OFDMA systems have been studied a lot recently. Such problems are generally multi-modal, non-convex and high-dimensional [11], [12]. Nonlinear programming algorithms are very popular for solving these problems as they converges very quick to fulfill real time service [13], [14], [15]. Nonconvex model is usually transformed to convex one based on some relaxation [16], [17], [18] or regularization techniques [19]. Many researches have been reported recently to produce more efficient algorithms [20], [21], [22], [23] and parallel execution of algorithms [24], [25]. Cicco et al. reported a video control plane which had performance prediction and regression method for resource allocation [26].

Metaheuristic algorithms are relatively new compared with nonlinear programming [27], [28]. Though they cost a large number of iterations to converge, these approaches tend to draw researchers attentions in wireless communication network designs [29]. Some of the famous metaheuristic algorithms are genetic algorithm (GA) [30], particle swarm optimization (PSO) [31], differential evolution (DE) [32], neighborhood field optimization [33] and artificial bee colony (ABC) algorithm [34] and so on.

DE is a simple yet easy to use optimization method [35]. Since its creation, it has been comprehensively applied to real world applications [36]. Most of them are modifications of classical DE, while these variants are more effective in solving given problem than classical method. This paper proposes to control algorithmic parameters in terms of wave propagation manner. Three waveforms are employed, i.e., sawtooth wave, square wave and triangle wave. The method can relieve the efforts to tune parameters for resource allocation problem. In our model, both network utility and fairness of all users are considered to maximize network throughput and assure high quality of service (QoS).

Section 2 briefly introduces resource allocation in wireless communication and gives the problem model. Section 3 presents the proposed algorithm and its analysis against existing methods. Section 4 gives simulation results compared with other algorithms. Resource allocation problem is also discussed in this section. Section 5 concludes the paper.