A Memetic Algorithm for Cache-aided Data Broadcast with Network Coding in Vehicular Networks Kai Liu ∗ Liang Feng ∗ Penglin Dai † Weiwei Wu ‡ ⋄ Victor C.S. Lee § Sang H. Son ¶ ∗ College of Computer Science, Chongqing University, Chongqing, China Email: liukai0807@cqu.edu.cn, liangf@cqu.edu.cn † School of Information Science and Technology, Southwest Jiaotong University, Chengdu, China Email: penglindai@swjtu.edu.cn ‡ School of Computer Science and Engineering, Southeast University, Nanjing, China Email:wweiwei2@gmail.com § Department of Computer Science, City University of Hong Kong, Kowloon, Hong Kong Email:csvlee@cityu.edu.hk ¶ Department of Information and Communication Engineering, DGIST, Daegu, Korea Email:son@dgist.ac.kr Abstract— With recent advances in wireless communications, vehicular networks are envisioned as a promising paradigm on achieving breakthroughs in transportation safety, efficiency, and sustainability. This work investigates data broadcast via Infrastructure-to-Vehicle (I2V) communication by exploiting the vehicular caching and network coding for enhancing band- width efficiency of the road-side unit (RSU). Specifically, we present an architecture for providing real-time data services via I2V communication in the service range of a RSU. Then, we investigate the problem of cache-aided data dissemination with network coding and prove that it is NP-hard. Further, we propose a memetic algorithm, which consists of a binary vector representation for encoding solutions, a fitness function for solution evaluation, a set of operators for offspring generation, a local search method for solution enhancement and a repair operator for fixing infeasible solutions. Finally, we build the simulation model and give a comprehensive performance eval- uation to demonstrate the superiority of the proposed solution. I. I NTRODUCTION Efficient data dissemination in vehicular networks is a key enabler for implementing emerging intelligent transportation systems (ITS), such as intersection control [1], platooning [2], accurate positioning [3], etc. However, intrinsic char- acteristics of vehicular networks including highly dynamic network topology, high mobility of vehicles, diverse and real-time information demands, etc., make it challenging to enhance system performance on information services by optimizing the utilization of the limited wireless bandwidth. The dedicated short-range communication (DSRC) is be- ing standardized as a defacto protocol in vehicular networks to support both infrastructure-to-vehicle (I2V) and vehicle- to-vehicle (V2V) communications [4]. A number of studies have proposed scheduling algorithms executed at the road- side unit (RSU) for data services via I2V communication [5], [6], [7], [8]. However, due to the the short I2V com- munication range, the limited wireless bandwidth and the ⋄ Corresponding author sparse deployment of RSUs, these approaches cannot best exploit the benefit of network resources to satisfy ever- increasing data service demands by ITS applications. Many studies considered to enhance communication quality and data throughput at the MAC layer. J. Zhang et al. [9] proposed a VC-MAC (Vehicular Cooperative Media Access Control) protocol, which jointly exploits the cooperative of I2V and V2V communications to enhance spatial reusability. The proposed protocol enables vehicles to cooperatively share their cached information via V2V communication when they are out of the RSU’s coverage. Y. Bi et al. [10] proposed a Multi-Channel Token Ring MAC Protocol (MC- TRP) for V2V communication. The CSMA/CA mechanism is applied for delivering emergency messages with low delay. In addition, a token-based data exchange protocol is designed to improve bandwidth efficiency for non-safety applications. From the application point of view, a number of studies have incorporated network coding into data dissemination in vehicular networks. M. Li et al. [11] investigated popular content distribution via push-based broadcast and proposed a cooperative content distribution protocol called CodeOn. The symbol level network coding technique is adopted to facilitate high speed content downloading and counteract the packet loss problem. W. Zhu et al. [12] proposed a multiple- vehicle protocol for collaborative data downloading using network coding. A multiple-vehicle collaborative download model is presented, and a probability mass function of the downloading completion time is derived to quantify the per- formance gains obtained from network coding. Our previous work [13] considered both communication constraints and application requirements in vehicular networks. A network coding-assisted scheduling algorithm is proposed to best exploit the joint effect of V2V and V2I communications and provide efficient data services. Distinguishing from previous efforts, this work aims to exploit the synergistic effect of network coding and vehicular caching for RSU-based data dissemination by allowing vehi- 978-1-5090-5019-2/17/$31.00 ©2017 IEEE