A Competitive and Dynamic Pricing Model for Secondary Users in - PowerPoint PPT Presentation

Outline A Competitive and Dynamic Pricing Model for Secondary Users in Infrastructure based Networks Soumitra Dixit, Shalini Periyalwar, and Halim Yanikomeroglu Broadband Communications and Wireless Systems (BCWS) Centre, Department of Systems and Computer Engineering, Carleton University, ON Canada September 08, 2010 Carleton University: S. Dixit, S. Periyalwar, H. Yanikomeroglu VTC Fall 2010 1/ 14 September 08, 2010

Introduction Inter-WSP Competition with Dynamic SU Pricing Outline Spectrum underutilization and Dynamic Spectrum Access (DSA) Distributed framework for Secondary User (SU) access Dynamic pricing model for SUs Multiple Wireless Service Providers (WSPs) and competitive SU pricing Achieving competitive yet dynamic SU pricing: Non-cooperative game theoretic analysis Dual benefits for SUs and WSPs Carleton University: S. Dixit, S. Periyalwar, H. Yanikomeroglu VTC Fall 2010 2/ 14 September 08, 2010

Overview Introduction System Concept Inter-WSP Competition with Dynamic SU Pricing SU Pricing Spectrum underutilization and DSA Spectrum occupancy field measurements: Underutilization of the radio spectrum in the spatial and temporal domains [Spectrum measurements, M. A. McHenry et al., ’06]. Dynamic Spectrum Access (DSA): Intelligent and efficient use of the radio spectrum by allowing opportunistic SU( unsubscribed ) access. Software Defined Radios (SDRs) or Cognitive Radios (CRs): envisioned to be enablers for DSA with the ability for cognition and reconfigurability. For infrastructure based networks: Potential for WSPs to gain additional profits by providing access to SUs. Carleton University: S. Dixit, S. Periyalwar, H. Yanikomeroglu VTC Fall 2010 3/ 14 September 08, 2010

Overview Introduction System Concept Inter-WSP Competition with Dynamic SU Pricing SU Pricing Previous Works Focus on centralized system framework with a Centralized Mediating Entity (CME) acting as a spectrum manager/broker/negotiator to pool the spectrum and manage the exchange of spectrum among WSPs and to SUs [Spectrum pooling: T. Weiss and F. Jondral ’04]. Dimsumnet architecture: Co-ordinated access band (spectrum pool) with ’spectrum broker’ for spectrum management [M. Buddhikot et al. ’05]. Spectrum Policy Server (SPS): negotiate spectrum on behalf of WSPs to SUs [O. Ileri et al. ’05]. Cognitive Pilot Channel (CPC): CPC manager for information exchange [J. Perez-Romero et al. ’07]. Competitive SU pricing and microeconomic models: [D. Niyato, E. Hussein, ’07]. Carleton University: S. Dixit, S. Periyalwar, H. Yanikomeroglu VTC Fall 2010 4/ 14 September 08, 2010

Overview Introduction System Concept Inter-WSP Competition with Dynamic SU Pricing SU Pricing Distributed System Framework Distributed Framework: Base Station(BSs) and not Wireless Service Providers(WSPs) individually advertise and sell their local unutilized spectrum to Secondary Users (SUs) [S. Dixit, S. Periyalwar, H. Yanikomeroglu, ’09]. Harmonious operation of Primary Users (PUs) and SUs at the same BS at equivalent power levels on different frequencies. Prioritized PU-SU scheduling: SU service subject to instantaneous spectrum availability after PUs have been served. SUs provided the freedom to select their preferred BS based on variety of parameters (i.e., price/service class, signal strength). Dynamic pricing model: SU price depends on spectrum resources utilized at the BS by its subscribers, i.e., PUs. Carleton University: S. Dixit, S. Periyalwar, H. Yanikomeroglu VTC Fall 2010 5/ 14 September 08, 2010

Overview Introduction System Concept Inter-WSP Competition with Dynamic SU Pricing SU Pricing Distributed Framework and Network Scenario ����������������������������� �������������������������� ����� �� ��� �������������������������� ������������������������������ � ����������� ���� � �� � ������ � ������� ��������� ����������� ��������� ��������� ���������� �������������� �� �� � ������ � ���������� �� � ������ � ���� ������� �� � ������� � ������� ����������� ���� Network scenario with a SU Snapshot of current spectrum requesting temporary wireless utilization at a particular BS. access from the BSs in the area. Carleton University: S. Dixit, S. Periyalwar, H. Yanikomeroglu VTC Fall 2010 6/ 14 September 08, 2010

Overview Introduction System Concept Inter-WSP Competition with Dynamic SU Pricing SU Pricing Dynamic Nature of SU Pricing: Terminology 4 s i : m i = 1 Inherent SU admission control s i → ∞ s i ’ : m i = 0.3 when α i,t → α i,th 3.5 Fixed PU price (per PU): p i = 1 Fixed cost (per PU/SU): c i = 0.45 3 Spectrum at BS i Spectrum currently available for SU access at BS i Spectrum at BS i SU price (per SU) s i 2.5 reserved currently occupied by PUs ( α i,pu , α i,th ) for handoff α i,pu = 0.4 and overload 2 protection Spectrum at BS i α i,h =0.1 currently available with 1.5 monetary incentive to the SUs ( α i,pu , α i,ic ) 1 0.5 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 SUF ( α i,t ) Carleton University: S. Dixit, S. Periyalwar, H. Yanikomeroglu VTC Fall 2010 7/ 14 September 08, 2010

Overview Introduction System Concept Inter-WSP Competition with Dynamic SU Pricing SU Pricing Dynamic Incentive based SU Pricing Model α i , h : Spectrum reserved for handoff; α i , th = 1 − α i , h . α i , su : Spectrum at BS i occupied by SUs; α i , su iff α i , pu < α i , th . α i , t : Spectrum Utilization Factor (SUF); α i , t = α i , pu + α i , su . α i , ic : Incentive cutoff limit beyond which s i , j > p i , j . SU Price ( s i ) w.r.t. PU price ( p i ) and SUF ( α i , t ) at the BS s i = ( f i ( α i , t )) m i × p i , ¯ (1) where s i , p i , ( f i ( α i , t ) , m i are non negative real numbers. m i : Price Leveling Factor (PLF) - additional pricing flexibility. Normalized SU price � α i , t � n i � � � − ln 1 − if 0 ≤ α i , t < α i , th , , f i ( α i , t ) = (2) α i , th ∞ , if α i , th ≤ α i , t ≤ 1 . Carleton University: S. Dixit, S. Periyalwar, H. Yanikomeroglu VTC Fall 2010 8/ 14 September 08, 2010

Multiple WSP Scenario Introduction Non-cooperative Game Theoretic Analysis Inter-WSP Competition with Dynamic SU Pricing Performance Results Competitive Pricing among Multiple WSPs Achieving competitive pricing 5 f 1,j ( α 1,t ) α 1,th = 0.9 4.5 f 2,j ( α 2,t ) with dynamic SU prices: reference PU price p i,j = 1 4 c i,j = 0.3 + 0.2( α i,t ) prohibitively complex. α 2,ic = 0.85 Normalized SU price f i,j ( α i,t ) 3.5 Current PU demand at For competitive dynamic pricing: BS of WSP 1 and WSP 2 3 2.5 α 1,pu = α 2,pu = 0.36 1) Use static SU prices ( S i ). 2 2) Find equilibrium static SU 1.5 α 1,ic = 0.7 α 2,ic = 0.75 prices. 1 3) Implement on dynamic model 0.5 0 using PLF( m i ). 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 SUF ( α i,t ) Tools (in step 2): Non-cooperative game theoretic Multiple WSPs: Aim to maximize analysis with SU service based individual WSP profits from SUs, differentiation. while competing on prices. Carleton University: S. Dixit, S. Periyalwar, H. Yanikomeroglu VTC Fall 2010 9/ 14 September 08, 2010

Multiple WSP Scenario Introduction Non-cooperative Game Theoretic Analysis Inter-WSP Competition with Dynamic SU Pricing Performance Results Two WSPs and the Differentiation of SU service �������� � �������� � ����������������� � � ��� ��� ��� ��� �� � ���� ��� �� �� � Identical service: high competition, low or zero profits. Differentiation of service: low competition, higher profits. Differentiation of the SU wireless service: using Dissatisfaction Price ( ζ ) based on the variance of the wireless channel ( σ i ); � σ 1 + σ 2 � ζ = K 1 K 2 ($), where K 1 = 1 ($); K 2 = . 2 Perceived price to each SU: U i ( y ) = S i + ( ζ × y ) ($). Carleton University: S. Dixit, S. Periyalwar, H. Yanikomeroglu VTC Fall 2010 10/ 14 September 08, 2010

Multiple WSP Scenario Introduction Non-cooperative Game Theoretic Analysis Inter-WSP Competition with Dynamic SU Pricing Performance Results Transformation for Achieving Competitive Pricing Nash Equilibrium (NE) SU price S ∗ i S ∗ i = C i + ζ, (3) where C i is the fixed cost considering static SU pricing. SU Pricing w.r.t. PU price and SUF at the BS i = ( f i ( α i , t )) m i × p i . s ′ (4) Mapping: s ′ i = S ∗ i , i.e., Static SU price mapped to first SU entering the BS at α i , t = α i , pu . Value of m i � S ∗ � ln i p i m i = (5) ln( f i ( α i , pu )) . Carleton University: S. Dixit, S. Periyalwar, H. Yanikomeroglu VTC Fall 2010 11/ 14 September 08, 2010