332 978-1-5090-4749-9/17/$31.00 ©2017 IEEE ICUFN 2017 A MOGA-Markov Chain Optimized Ranking Algorithm for Wireless Access Networks in Heterogeneous Environment Qazi Zia Ullah 1,2 , Farman Ullah 1 , Fazal Wahab Karam 1 , Shahzad Hassan 2 , Sungchang Lee 3 1 Electrical Engineering Department, 2 Computer Engineering Department, CIIT Attock Campus, Pakistan Bahria University Islamabad, Pakistan 3 School of Electronics and Information Engineering, Korea Aerospace University, South Korea Abstract – The mounting customer demands for bandwidth-desirous services are deriving for a cost effective, robust, and high capacity wireless access network. The end users expect a satisfactory and economical delivery of “Quad-play” applications (voice, video, data, and mobility) and rich-media applications (multimedia, interactive gaming, and meta-verse) over the wireless network. In last two decades, a remarkable evolution of wireless networks is observed. Moreover, after the advent of software defined radio enabled wireless sets, the selection of the optimum wireless access network for different applications (live video streaming, online gaming, voice calling and browsing) is gaining vital importance. In this paper, a ranking algorithm based on Markov chain optimized learning approach is formulated for the heterogeneous environment. The algorithm is designed on the basis of most important Quality of Service (QoS) parameters like throughput, delay/error and cost. The proposed technique is robust against the change in number of available networks where, previously proposed techniques TOPSIS, VIKOR and RafoQ are unable to handle the change in available networks adequately. The Simulation results verify the selection of optimal access network for varying applications conforming to defined ranking algorithm. Keywords – Ranking; Wireless access network; Optimized learning; Heterogeneous networks; Quality of service (QoS). INTRODUCTION The wireless fidelity (Wi-Fi), long term evolution (LTE) and worldwide interoperability for microwave access (WiMAX) are emerging as the key technologies for heterogeneous wireless environment as depicted in fig-1. In previous studies, different cellular networks like UMTS, GSM, GPRS, HSPA and different standards of WLAN were simultaneously studied for designing the ranking algorithm. In this paper, the discussion is restricted to LTE, WiMAX and Wi-Fi, as cellular networks are converging towards LTE and WLANs towards different flavors of Wi-Fi. The Wi-Fi is defined as any wireless local area network (WLAN) based on the Institute of Electrical and Electronics Engineers (IEEE) 802.11 standards. Hence, all 802.11 networks are adverted as Wi-Fi, like 802.11a, 802.11b, 802.11g and 802.11n [1]. 802.11a works at 5GHz frequency band, whereas 802.11b and 802.11g operate at 2.4GHz frequency band [2]-[3]. While on the contrary, LTE is designed to exchange high speed data among cell phones and data-terminals. It is egressed from GSM, EDGE, UMTS and HSPA network technologies but raising the network capacity and data rate many folds by improving the radio interface [4]-[5]. BTS BTS BTS SU Fig.1. Heterogeneous network scenario WiMAX is designed to offer 30 to 40 Mbit/s data rates for broadband access but with the 2011 amendment is able to offer 1Gbit/s for fixed wireless stations [6]. WiMAX is based on the two IEEE standards 802.16e and 802.16m. The 802.16e specifies the point to multipoint fixed wireless communications to deliver last-mile wireless broadband service as a substitute to cable and digital subscriber line (DSL). However, the 802.16m is an amendment to 802.16e to enable mobile usage in cellular networks by providing efficient handover techniques [7]. The 3GPP and IEEE have released many variations of LTE, WiMAX, and Wi-Fi, which supports different data rates. The choice of most suitable wireless access network for varying nature of applications requires some ranking criterion. Moreover, the performance parameters are distinctly different for applications, such as video streaming, browsing and voice calling. They are quite dissimilar in signal generation, propagation and reception, such as, the voice/video calling is a delay sensitive service and requires a dedicated end-to-end connectivity. On the other hand, browsing is not delay sensitive and works on the best delivery mechanism. Therefore, only one network will not be able to fulfill these diversifying requirements. Moreover, installation, working, and maintenance of wireless access networks differ in design, cost and structure. To choose the best access network for a particular application, an analysis of different QoS parameters need to perform. The ranking is not possible by considering only a single parameter. So, the complete study and analysis is required for suitability of an access network for a specific application by considering the broad categories of parameters