Optimal Resource Allocation For Multi-tier Heterogeneous Networks

The main task of resource allocation is to meet the demands of multiple users.That is,provide quality of services.Resource allocation directly determines the communication performance of users.Due to the relative scarcity of spectrum resources,and the random nature of the wireless channel,a lot of work has been concerned the resource allocation in the past decades.In the near future,appearing applications from smart-city and machineto-machine sharpen the conflict between the limited radio resource and the rapid growth of users' requirements.Small-cell heterogeneous networks helps the Macro-cell base station offloading the data services,by increasing the density of access nodes.Small-cell heterogeneous networks have advantages of flexibility,high efficiency.Both scholars and engineers confirm that,multi-tier small-cell heterogenouse network is one of the most promising solution.Consequently,it is of great significance for us to study the resource allocation for small-cell heterogeneous networks.We first research the small-cell BS assignment and BS sleeping.We view the small-cell network base stations(BSs)as one type of spatial resource.The BSs are allocated according to the demand of the users.In green cellular networks,small-cell base stations can be allocated under the control of resource management entities.The unassigned BSs are asleep.Assume there are total K active users are co-channel accessing.We establish the small-cell BS assignment problem as an mixed integer programming problem.The complexity of this problem will be exponential increase with the number of small-cell BSs and user K.To trackle the complexity,we derive the resource allocation algorithm having polynomial-time complexity.Furthermore,we design the distributed implementation of these two algorithms.Wether activate one small-cell BS depends on the results of the resource allocation.The conventional rule is to assign small-cell BS to users one-to-one,one user only needs one smallcell BS.In our algorithms,whether the allocation of one small-cell base station to the active user depends on two aspects: 1.whether the user's basic requirement is satisfied;2.whether the benefits from diversity gain is over the network overhead consumption.The simulation results are analyzed based on power cost and also computational complexity.One important conclusion is that: the tradeoff between the user's transmitting power and the power needed for the small-cell BS can be obtained.Then,the limit performance of resource allocation is studied.The types of resource considered here are: power,multi-antennas and small-cell BS nodes.There are infeasible cases in the above algorithm we proposed if there are not enough aviable small-cell BSs.Therefore,we need to understand the achieveable performance boudary.With limited power resource,we characterize the reachable performance bounds for multi-users.As long as one user's QoS requirement exceeds the performance boundary,the small-cell network is not feasible.We use the Pareto optimal boundary to characterize the boundary of this region.The traditional characterization of Pareto boundary has not contained the BS assignment.Viewing the small-cell BS nodes as one kind of resource,we make progress on characterizing the Pareto boundary.However,cooperative multi-point reception will bring more complexity to the characeterization.To this end,we propose two kinds of low complexity strategies: network centric and user centric allocation strategy.The network centric allocation refers to select the BS set,and then notify the user to access.THe user centric strategy assume that the user has permissions to wake up the small-cell BS,and then access by itself.Numerical experiments are carried out.First,we verify the accuracy of our method according to comprison with the classical benchmark.Next,we characterize the Pareto optimal boundary including the base station node resource,which has additional gain compared to the classical benchmark.Considering the complexity and performance,the advantages and disadvantages of two small-cell base station node allocation strategies are clearly presented by experiments.Multiuser interference is an unfavorable factor in the traditional communication system.However,it can be turned into a protecting resource for confidential message.Deploying at premises of end-users,small-cell base stations are in the insecure environments.The transmitters have a large number of private message,such as account and password.Considering the co-existence of normal users and secure users,guarantee the secrecy is an improtant problem.Our method is: using secure power allocation and precoder to provide physical layer secrecy.We consider multiple normal users as the potential eavesdroppers.Guaranteeing the quality of the security service,we model two kinds of quality of services constraints:instantaneous SINR threshold and long-term outage probability constraint.Compared to the traditional QoS-based multi-cell beamforming problem,our problem has additional secrecy constraints,becomes non-convex.We use semidefinite relaxation and convex-concave approximation procedure to solve these two problems.In different parameters we make numerical examples to analyze.The numerical results demonstrate the effecacy of our proposed method.It is concluded that even there are multiple potential eavesdroppers,we can achieve guaranteeing the quality of secrecy services.