Dynamic Decision-Based Spectrum Sharing Framework for Next-generation (5G) Systems Zhaleh Sadreddini 1 , Pavel Masek 2 , Tugrul Cavdar 3 , Jiri Hosek 2 , and Erkan Guler 4 1 Faculty of Communication, Giresun University, Giresun, Turkey 2 Department of Telecommunications, Brno University of Technology, Brno, Czech Republic 3 Department of Computer Engineering, Karadeniz Technical University, Trabzon, Turkey 4 Department of Computer Engineering, Giresun University, Giresun, Turkey Contact author’s e-mails: zh.sadreddini@ktu.edu.tr; masekpavel@feec.vutbr.cz Abstract—Looking into the concept of next-generation (5G) cellular systems, it is necessary to do a revision of existing radio spectrum management techniques and come up with more flexible solutions. A new wave of spectrum policy reforms can be envisaged with a direction shift from static to dynamic optimiza- tion. According to the peak hours, the number of served users in mobile networks is increasing. Since the radio spectrum is limited, cognitive radio (CR) technology provides an opportunity to recognize under-utilized cellular spectrum (licensed band) resources. To this end, efficient spectrum management techniques based on CR technology should be implemented to share the spectrum between different types of users in order to maximize spectrum utilization and spectral efficiency. In this work, we present dynamic decision-based spectrum sharing model among multiple classes of users in CR network (CRN) in order to increase network utilization and the quality of experience (QoE) by increasing the users’ satisfaction. Obtained simulation results from created toolkit in Matlab tool (calibrated by data set from real 3GGP LTE-Advanced system) show the performance of the developed model and appropriate user selection among multiple users’ types. Index Terms—Cognitive Radio Network, Spectrum Manage- ment, Spectral Efficiency, User Satisfaction, Next-generation (5G) Cellular Systems. I. I NTRODUCTION Over the last decades, radio spectrum has transformed into a critical resource from the economic, cultural and social point of view. Since the spectrum scarcity has been proven to be a major issue across particular frequency ranges, spanning from 100 MHz to 6 GHz, the need for advanced spectrum sharing between limited number of users while guaranteeing their interference protection is expected to play crucial role [1], [2]. However, cellular spectrum resources are not being utilized by primary user (PU) at a specific time and location. To make the spectrum utilization more efficient, a secondary user (SU) can be allowed to access under-utilized cellular spectrum at the given time instant and geographic location 1 . This type of dynamic spectrum allocation can be done utilizing ”Cognitive Radio” which uses a Software Defined Radio (SDR) principles by efficient allocation of under-utilized resources (UR) by SUs as long as the primary activity remains idle [3], [4]. 1 Dynamic decision-based spectrum sharing framework enables spectrum sharing by allowing at least two users, the PU (i.e., current holder of spectrum rights (mobile operator)) and SU (i.e., temporary user of spectrum). 978-1-5090-6494-6/17/$31.00 c 2017 IEEE CRN includes four main functional blocks such as: (i) Spec- trum Sensing, (ii) Management, (iii) Sharing, and (iv) Mo- bility. All CRN functional blocks are handled by network operator (NO) entity which is demonstrated as Decision- maker. However, according to increasing number of connected users, the time given to NO for decision regarding spectrum allocation is decreasing as the number of requesting SUs (RSUs) grows. To this end, systematic framework based on a scientific background is needed to make appropriate decisions. The problem can be abstracted as a question, how to derive weights, rankings or importance ratio for a set of activities according to their impact on the situation in mobile network. This approach is known as the process of Multiple-Criteria Decision-Making (MCDM) [5]. Next, the Analytic Hierarchy Process (AHP) is a structured MCDM technique for organizing and analyzing complex decisions [5], [6], [7]. It enables a particular application in decision making and is widely used in variety of decision-making scenarios (i.e., prioritization, resource allocation, benchmarking, and quality management). In accordance to the limited cellular spectrum resources, dynamic spectrum sharing becomes a key enabler for maximal spectrum utilization and spectral efficiency, especially during peak hours. Therefore, demand for an optimal decision-based spectrum management and charging policies take place these days. In [8], authors propose two centralized optimization frame- works for maintaining Quality of Service (QoS) in a multi- channel CRN. Authors in [9] introduce fairness weights for each user – capturing current and past data sets of user’s experience. Further, authors design evolution models for the fairness weights based on society model. A game theoretic approach for resource allocation in case of primary and secondary users in cognitive networks is presented in [10]. Further, in [11], a decision algorithm for spectrum brokers is proposed for heterogeneous CRNs considering AHP method. Adaptive decision-making scheme for CRN with multiple sub- carriers is further presented in [12]. Above mentioned works are related to the spectrum management, nevertheless, none of them pay attention to Quality of Experience (QoE). In this paper, we focus on the situation, where NO uses AHP method to allocate an optimal importance degree for RSUs – in order to provide service to the appropriate RSUs and to achieve not only the highest level of QoS, but also improve