CN106358230A - Load balancing method based on heterogeneous network - Google Patents

Abstract

The invention relates to heterogeneous network load balancing, in particular to a method for balancing traffic load levels between base stations (BS) and between a BS and an access point (AP) through a suitable method by using interaction between a wireless access point (AP) and a base station (BS) of a mobile communication network. The technical scheme adopted by the invention is that the load balancing method based on a heterogeneous network comprises the following steps of: S1, performing traffic balancing among the base stations, wherein a basic relational expression of a system is as shown in the specification; and S2, performing traffic unloading from the base station to the access point. The method provided by the invention is mainly applied to occasions of heterogeneous network load balancing.

Description

Load-balancing method based on heterogeneous network

Technical field

The present invention relates to heterogeneous network load balancing, especially for the answering of multi-access point in macro base station and its coverage Close the load balancing in heterogeneous network.

Background technology

Heterogeneous network is current and the major networks structure of next generation mobile communication, it be Wireless Personal Network (such as Bluetooth), WLAN (as wi-fi), wireless MAN (as wimax), public mobile network (as 3g, 4g) with And the heterogeneous wireless network that multiple access networks such as adhoc network coexist.And the development with mobile communication, user is to wireless data The demand of flow can be increasing, and the single unit (as public mobile network network) in heterogeneous network is easy to bear larger Data traffic pressure, needs to find a kind of method, using the characteristic of heterogeneous network multi-level-cell, the network of overload is born Carry balance and flow unloading.

Content of the invention

For overcoming the deficiencies in the prior art, it is contemplated that by using WAP (access point, ap) with Interaction between base station of mobile communication network (base station, bs), balance between bs with a kind of suitable method and bs with Flow load level between ap.The technical solution used in the present invention is, based on the load-balancing method of heterogeneous network, step is such as Under:

The first step is the flow equilibrium between base station, and the basic relational expression of system is

$\stackrel{\·}{{\mathrm{\θ}}_x}={\mathrm{\ω}}_x+\mathrm{\γ}\underset{j=1}{\overset{n}{\σ}}sin({\mathrm{\θ}}_y-{\mathrm{\θ}}_x),x=1,...,n$

Wherein n is phase coupling estimation oscillator quantity, θ_xAnd θ_yFor the phase place of oscillator x and y, ω_xFor the frequency of oscillator x, γ is coupling Syzygy number；

If the load level after the normalization of base station be ρ ∈ (0,1], then for base station x, the load level of next time slot should For

ρ_x(t+1)=ρ_x(t)+δ_x(t)

Wherein ρ_xT () is the load level of t base station x, δ_xRequired for the x of base station, the load capacity of transfer, is defined as

${\mathrm{\δ}}_x=\mathrm{\γ}*\underset{y\&element;n_x}{\σ}sin({\mathrm{\ρ}}_y-{\mathrm{\ρ}}_x)$

Wherein n_xIt is the set with base station x adjacent base station, ρ_xWith ρ_yIt is respectively the load level of base station x and base station y；

Second step is the flow unloading to access point for the base station: sets u (ρ) as quantifying the utility function of Consumer's Experience, is defined as

$u\left(\mathrm{\ρ}\right)=b\sqrt{1-{\mathrm{\ρ}}^2}$

Wherein b is value of utility when 0 it is assumed that all there being several wifi to access in each base station range for load of base station Point, when base station unloads, to these access points, the flow that total amount is σ, base station can produce a utility function increment, that is,

δ u (σ)=u (ρ-σ)-u (ρ)

And when an access point receives the flow unloading the total amount being downloaded from base station for ε, access point can bear certain pressure Power, is defined as:

δ c (ε)=β * ε

Wherein β is the pressure coefficient of this access point, it to and being directly proportional apart from d of base station；

In flow uninstall process, need the difference maximum making base station effective increment and accessing point pressure, then base station is to access The flow unloading problem of point can be converted into a convex optimization problem, that is,

$\underset{{\mathrm{ap}}_i\&element;m_p}{min}\mathrm{\σ}\mathrm{\δ}c\left({\mathrm{\ϵ}}_{{\mathrm{ap}}_i}\right)-{\mathrm{\δu}}_p\left(\mathrm{\σ}\right),$

S.t. σ=∑ ε

${\mathrm{\ϵ}}_{{\mathrm{ap}}_i}\&greaterequal;0,{\mathrm{ap}}_i\&element;m_p$

Wherein p is base station sequence number, m_pFor the set of access point in the p coverage of base station,It is to unload to i-th access point The flow carrying.

The feature of the present invention and beneficial effect are:

This programme has multiple base stations and the heterogeneous network of numerous access point covering to emulate by emulation mode to one, Scheme is implemented, and can quickly be reduced the load level of base station.

Brief description:

Fig. 1 heterogeneous network schematic diagram,

Flow equilibrium between Fig. 2 base station,

Fig. 3 utility function image,

Fig. 4 base station is to the flow unloading of access point.

Specific embodiment

The model that the present invention adopts includes following two:

1) kuramoto Vibration Synchronization model: it is located in a synchronization system and have two oscillator a and b, their frequency phase With phase place is respectively θ₁And θ₂, then phase shift after they interact in synchronizing process

δ=γ * sin (θ₁-θ₂)

Wherein γ is the coefficient of coup of system.

In the present invention, Vibration Synchronization model is used for carrying out load balance between two base stations.

2) utility function optimizing model: utility function be conventionally used to indicate consumer consumption obtained in effectiveness with The function of quantitative relation between the grouping of commodities consumed is full obtained in set grouping of commodities from consuming to weigh consumer The degree of foot.Represent the current consumption experience of user in the present invention by arranging base station based on the utility function of load level, Pressure produced by flow is received by the difference of the value of utility before and after flow is unloaded and access point and carries out optimization, obtain outflow Unloading scheme.

Below in conjunction with the accompanying drawings and instantiation, further describe the present invention.

The load balancing that algorithm is divided between base station and base station unload two steps to the flow of access point, respectively in connection with Kuramoto synchronistic model and the utility function model based on Consumer's Experience.

The first step is the flow equilibrium between base station, as shown in Figure 2.Synchronization system (as neutral net) can regard an arteries and veins as Rush coupled oscillationses network.Agitator adjusts the vibration of itself so that system tends to same by observing the attribute of other oscillators Step.Kuramoto is exactly a typical phase coupling estimation oscillatory system model.In this system, each oscillator is had by oneself with it Frequency Independent Vibration, and tend to affect other oscillators with this.The basic relational expression of system is

$\stackrel{\·}{{\mathrm{\θ}}_x}={\mathrm{\ω}}_x+\mathrm{\γ}\underset{j=1}{\overset{n}{\σ}}sin({\mathrm{\θ}}_y-{\mathrm{\θ}}_x),x=1,...,n$

Wherein n is phase coupling estimation oscillator quantity, θ_xAnd θ_yFor the phase place of oscillator x and y, ω_xFor the frequency of oscillator x, γ is coupling Syzygy number.

If the load level after the normalization of base station be ρ ∈ (0,1], then for base station x, the load level of next time slot should For

ρ_x(t+1)=ρ_x(t)+δ_x(t)

Wherein ρ_xT () is the load level of t base station x, δ_xRequired for the x of base station, the load capacity of transfer, is defined as

${\mathrm{\δ}}_x=\mathrm{\γ}*\underset{y\&element;n_x}{\σ}sin({\mathrm{\ρ}}_y-{\mathrm{\ρ}}_x)$

Wherein n_xIt is the set with base station x adjacent base station, ρ_xWith ρ_yIt is respectively the load level of base station x and base station y.

Second step is the flow unloading to access point for the base station.Utility function is conventionally used to indicate consumer and is obtained in consumption The function of quantitative relation between the effectiveness obtaining and the grouping of commodities consumed, to weigh consumer from the set grouping of commodities of consumption Obtained in meet degree.For each base station, Consumer's Experience (including data rate, stability etc.) is always born with base station It is loaded with pass, if u (ρ) is the utility function quantifying Consumer's Experience, be defined as

$u\left(\mathrm{\ρ}\right)=b\sqrt{1-{\mathrm{\ρ}}^2}$

Wherein b is the value of utility that load of base station ρ is when 0, functional image such as Fig. 3.

As Fig. 4 it is assumed that all there being several wifi access points in each base station range, when base station unloads to these access points When total amount is the flow of σ, base station can produce a utility function increment, that is,

δ u (σ)=u (ρ-σ)-u (ρ)

And when an access point receives the flow unloading the total amount being downloaded from base station for ε, access point can bear certain pressure Power, is defined as

δ c (ε)=β * ε

Wherein β is the pressure coefficient of this access point, it to and being directly proportional apart from d of base station.

In flow uninstall process, need the difference maximum making base station effective increment and accessing point pressure, then base station is to access The flow unloading problem of point can be converted into a convex optimization problem, that is,

$\underset{{\mathrm{ap}}_i\&element;m_p}{min}\mathrm{\σ}\mathrm{\δ}c\left({\mathrm{\ϵ}}_{{\mathrm{ap}}_i}\right)-{\mathrm{\δu}}_p\left(\mathrm{\σ}\right),$

S.t. σ=∑ ε

${\mathrm{\ϵ}}_{{\mathrm{ap}}_i}\&greaterequal;0,{\mathrm{ap}}_i\&element;m_p$

Wherein p is base station sequence number, m_pFor the set of access point in the p coverage of base station,It is to unload to i-th access point The flow carrying.

Pseudo-code of the algorithm is as shown above.

Claims (1)

1. a kind of load-balancing method based on heterogeneous network, is characterized in that, step is as follows:

The first step is the flow equilibrium between base station, and the basic relational expression of system is

${\stackrel{\·}{\mathrm{\θ}}}_x={\mathrm{\ω}}_x+\mathrm{\γ}\underset{j=1}{\overset{n}{\σ}}sin({\mathrm{\θ}}_y-{\mathrm{\θ}}_x),x=1,\mathrm{...},n$

Wherein n is phase coupling estimation oscillator quantity, θ_xAnd θ_yFor the phase place of oscillator x and y, ω_xFor the frequency of oscillator x, γ is coupled systemes Number；

If the load level after the normalization of base station be ρ ∈ (0,1], then for base station x, the load level of next time slot should be

ρ_x(t+1)=ρ_x(t)+δ_x(t)

Wherein ρ_xT () is the load level of t base station x, δ_xRequired for the x of base station, the load capacity of transfer, is defined as

${\mathrm{\δ}}_x=\mathrm{\γ}*\underset{y\&element;n_x}{\σ}sin({\mathrm{\ρ}}_y-{\mathrm{\ρ}}_x)$

Wherein n_xIt is the set with base station x adjacent base station, ρ_xWith ρ_yIt is respectively the load level of base station x and base station y；

Second step is the flow unloading to access point for the base station: sets u (ρ) as quantifying the utility function of Consumer's Experience, is defined as

$u\left(\mathrm{\ρ}\right)=b\sqrt{1-{\mathrm{\ρ}}^2}$

Wherein b for load of base station be value of utility when 0 it is assumed that all there being several wifi access points in each base station range, when During the flow that base station is σ to these access points unloading total amount, base station can produce a utility function increment, that is,

δ u (σ)=u (ρ-σ)-u (ρ)

And when an access point receives the flow unloading the total amount being downloaded from base station for ε, access point can bear certain pressure, fixed Justice is:

δ c (ε)=β * ε

Wherein β is the pressure coefficient of this access point, it to and being directly proportional apart from d of base station；

In flow uninstall process, need the difference maximum making base station effective increment and accessing point pressure, then base station is to access point Flow unloading problem can be converted into a convex optimization problem, that is,

$\underset{{\mathrm{ap}}_i\&element;m_p}{min}\mathrm{\σ}\mathrm{\δ}c\left({\mathrm{\ϵ}}_{{\mathrm{ap}}_i}\right)-{\mathrm{\δu}}_p\left(\mathrm{\σ}\right),$

S.t. σ=∑ ε

${\mathrm{\ϵ}}_{{\mathrm{ap}}_i}\&greaterequal;0,{\mathrm{ap}}_i\&element;m_p$

Wherein p is base station sequence number, m_pFor the set of access point in the p coverage of base station,It is to i-th access point unloading Flow.