CN102480736B - Method and device for configuring dynamic data service channel - Google Patents

Abstract

The invention provides a method and a device for configuring a dynamic data service channel. The configuration method comprises the following steps that: an allocation proportion of a dynamic packet data channel (PDCH) is introduced into a Markov model, a punishment entering and delay releasing mechanism of a data user wireless channel is introduced into the model, at the same time a current network half-rate proportion is used as a threshold for triggering the wireless channel punishment entering and delay releasing mechanism, a best dynamic PDCH configuration value of a wireless carrying efficiency under the restriction of the system channel performance is satisfied is calculated through the numeric value and the system emulation, so the wireless channel carrying efficiency of the system can be improved, the automatic configuration of the dynamic PDCH allocation proportion can be realized, the intelligentized and automatic adjustment of dynamic PDCH allocation proportion parameters of communities can be realized, direct allocation workload of network personnel can be alleviated, and the configuration accuracy can be improved.

Description

A kind of collocation method of dynamic data Traffic Channel and device

Technical field

The present invention relates to wireless communication technology field, particularly relate to collocation method and the device of dynamic data Traffic Channel (DPDCH) in GPRS (GeneralPacket Radio Service) system.

Background technology

In gprs system, the Traffic Channel of a community is generally jointly made up of voice service channel (TCH) and data traffic channels (PDCH).Data traffic channels comprises fixing PDCH (FPDCH, Fix PDCH) and dynamic PDCH (DPDCH, Dyn PDCH).Wherein Fix PDCH is only for bearing data service, can not be used for carry voice traffic.When the more PDCH of needs be used for bearing data service time, system using the TCH of free time as dynamic PDCH.When there being new audio call to arrive, shortage appears in TCH, dynamic PDCH will be cleared, and distribute to corresponding audio call as TCH.

For convenience of description, in the present patent application, define the maximum that dynamic PDCH allocation proportion is the ratio that can be converted to dynamic PDCH in TCH, be designated as Max_Dyn_Ratio (representing with η).When the dynamic PDCH quantity of having distributed in system reaches TCH × Max_Dyn_Ratio, by the new dynamic PDCH that do not reallocate.Voice-and-data traffic channel assignments schematic diagram is as shown in Figure 1:

For the setting of dynamic PDCH allocation proportion, existing technical scheme is with existing network performance statistics and basis, and the easy configuration in conjunction with empirical data and producer's suggestion plans the Configuration Values drawing corresponding device.Configuration needs to rely on optimization personnel manual modification plant equipment parameter.

For the strategy that fixing PDCH distributes, main technical scheme is the wireless channel distribution method based on markov (Markov) chain.Because GPRS/EDGE introduces on original GSM network basis, limited channel resource is shared in data service and speech business, can do not distributed to GPRS/EDGE dynamically as required by the channel that speech business takies, this dynamic process can describe with the Markov process of a moving boundary.In the method, setting TCH number is C, FPDCH number is D, and supposes that the arrival rate of speech business and data service all obeys Poisson (Poisson) distribution, and all obeys quantum condition entropy service time, and the waiting list of packet exists limiting value.

The method of the fixing PDCH distribution of prior art, also exists following problem:

1. prior art is based on the wireless channel distribution method of Markov chain, its research be the impact that fixing PDCH divides pairing network performance, for TCH, as long as then think the free time just can be assigned as PDCH.But in existing network community, TCH can be converted to the value of dynamic PDCH is limited, this ratio identifies by dynamic PDCH allocation proportion usually.In existing technical scheme, this key index of maximum allocated ratio of dynamic PDCH is not considered completely.But in fact this parameter limits the complexity that dynamic PDCH applies for, is far-reaching on the impact of systematic function.Therefore, the network performance that prior art calculates estimates the actual conditions that can not reflect network well.

2. supposition user is regardless of the grade of service when studying voice user and accessing for the existing wireless channel distribution method based on Markov chain, and all users one arrive and namely allow access.No matter therefore whether wireless channel is congested in a model, the arrival rate of user is consistent.In the equipment controlled there being the grade of service, will start congestion control when system occurs congested, that namely passes through the quick access attenuating low-end subscribers of refusal low-end subscribers reaches rate.On the other hand for data user, existing model assumption user has served and has left system.But in existing network equipment, leaving of data user is controlled by delay strategy, be not immediately leave.Therefore, prior art does not consider the actual conditions that existing network discharges for access control and time delay.

As the above analysis, the setting of dynamic PDCH allocation proportion (Max_Dyn_Ratio) value is a very crucial problem, the dynamic PDCH that this value arranges more small data traffic assignments is fewer, the degree that speech resources is fought in data service is less, effectively can reduce voice half rate ratio and TCH congestion ratio, promote every PDCH bearer traffic; But the too small distribution that will affect data service dynamic channel is set, affects the perception of data service customers.Therefore, being set in the setting parameter of system cell of dynamic PDCH allocation proportion is a very crucial value, needs the method finding the dynamic PDCH allocation proportion of a kind of reasonable disposition badly, to promote the radio bearer efficiency of network.

Summary of the invention

Embodiment of the present invention technical problem to be solved is to provide a kind of collocation method and device of dynamic data Traffic Channel, realizes the automatic distribution of dynamic PDCH in gprs system, promotes radio bearer efficiency.

For solving the problems of the technologies described above, the embodiment of the present invention provides scheme as follows:

A collocation method of dynamic PDCH, comprising:

According to user behavior statistics in community, calculate voice user in described community, average arrival rate that data user enters system and by the average service rate of system service;

For described community, according to average arrival rate and the average service rate of voice user, data user, the two-dimensional Markov chain that to set up with the data user's number in voice traffic channel number current in system, data user's waiting list be two-dimensional parameter;

According to the magnitude relationship of voice traffic channel number current in system and channel thresholding L, revise arrival rate and the service rate of data user in described two-dimensional Markov chain, wherein said channel thresholding L=N (1-δ), N is the total channel number of data and voice sharing in described community, and δ is the half rate ratio of voice channel in described community;

Under different dynamic PDCH allocation proportions, calculate the steady state solution of described two-dimensional Markov chain, obtain the state transition probability matrix that two-dimensional Markov chain is corresponding, and based on described state transition probability matrix, computing system performance parameter, and judge whether the system performance parameter calculated meets predetermined system performance requirements, when meeting described system performance requirements, calculates the load-carrying efficiency of the wireless channel under this dynamic PDCH allocation proportion;

Select optimum wireless channel load-carrying efficiency, by the dynamic PDCH allocation proportion corresponding to the load-carrying efficiency of the wireless channel of this optimum, as the allocation ratio of dynamic PDCH.

Preferably, in above-mentioned collocation method,

The two-dimensional Markov chain that described foundation is two-dimensional parameter with the data user's number in voice traffic channel number current in system, data user's waiting list, comprising:

For each the state P in two-dimensional Markov chain _{i, j}, the arrival rate λ of calculated data user ₁=(M-i) λ _dwith service rate μ _i(j)=min (N-j, D+ η C, i) μ _d, wherein, i=0,1 ..., M, represents the data user's number in data user's waiting list; J=0,1 ..., N, voice traffic channel number current in expression system, N represents the total channel number of data and voice sharing in described community, and M represents the data user's sum activated in described community, and C represents the TCH number of channel, D represents the FPDCH number of channel, and η represents dynamic PDCH allocation proportion, λ _dand μ _drespectively represent data user enter system average arrival rate and by the average service rate of system service;

With the arrival rate λ of described data user _iwith service rate μ _i(j), as the transition probability of data user's number increase and decrease in two-dimensional Markov chain, and, with the average arrival rate of described voice user and average service rate, as the transition probability of voice traffic channel number increase and decrease in two-dimensional Markov chain, set up two-dimensional Markov chain.

Preferably, in above-mentioned collocation method,

The described magnitude relationship according to described voice traffic channel number and channel thresholding L, revise arrival rate and the service rate of data user in described two-dimensional Markov chain, comprising:

When voice traffic channel number current is in systems in which more than or equal to described channel thresholding L, reduce the arrival rate of data user; When described voice traffic channel number is less than described channel thresholding L, reduce the service rate of data user.

Preferably, in above-mentioned collocation method,

The arrival rate of described reduction data user is specially: by the arrival rate of data user by λ _ibe reduced to λ _iq _j, wherein, q _jbe one and be less than 1 and the function reduced along with the increase of j;

The service rate of described reduction data user is specially: by the service rate of data user by μ _ij () is reduced to μ _i(j) k _j, wherein, k _jbe one and be less than 1 and the function reduced along with the increase of j.

Preferably, in above-mentioned collocation method,

Described system performance parameter comprises the probability of substitute of data service, the average queue length of data service, at least one in the blocking rate of the average delay of packet in queue and speech business calling.

The embodiment of the present invention additionally provides the inking device of a kind of dynamic PDCH, comprising:

Computing unit, for according to user behavior statistics in community, calculates voice user in described community, average arrival rate that data user enters system and by the average service rate of system service;

Modeling unit, for for described community, according to average arrival rate and the average service rate of voice user, data user, the two-dimensional Markov chain that to set up with the data user's number in voice traffic channel number current in system, data user's waiting list be two-dimensional parameter;

Amending unit, for the magnitude relationship according to voice traffic channel number current in system and channel thresholding L, revise arrival rate and the service rate of data user in described two-dimensional Markov chain, wherein said channel thresholding L=N (1-δ), N is the total channel number of data and voice sharing in described community, and δ is the half rate ratio of voice channel in described community;

Simulation unit, for under different dynamic PDCH allocation proportions, calculate the steady state solution of described two-dimensional Markov chain, obtain the state transition probability matrix that two-dimensional Markov chain is corresponding, and based on described state transition probability matrix, computing system performance parameter, and judge whether the system performance parameter calculated meets predetermined system performance requirements, when meeting described system performance requirements, calculate the load-carrying efficiency of the wireless channel under this dynamic PDCH allocation proportion;

Selected cell, for selecting optimum wireless channel load-carrying efficiency, by the dynamic PDCH allocation proportion corresponding to the load-carrying efficiency of the wireless channel of this optimum, as the allocation ratio of dynamic PDCH.

Preferably, in above-mentioned inking device,

Described modeling unit, specifically for:

For each the state P in two-dimensional Markov chain _{i, j}, the arrival rate λ of calculated data user _i=(M-i) λ _dwith service rate μ _i(j)=min (N-j, D+ η C, i) μ _d, wherein, i=0,1 ..., M, represents the data user's number in data user's waiting list; J=0,1 ..., N, voice traffic channel number current in expression system, N represents the total channel number of data and voice sharing in described community, and M represents the data user's sum activated in described community, and C represents the TCH number of channel, D represents the FPDCH number of channel, and η represents dynamic PDCH allocation proportion, λ _dand μ _drespectively represent data user enter system average arrival rate and by the average service rate of system service;

With the arrival rate λ of described data user _iwith service rate μ _i(j), as the transition probability of data user's number increase and decrease in two-dimensional Markov chain, and, with the average arrival rate of described voice user and average service rate, as the transition probability of voice traffic channel number increase and decrease in two-dimensional Markov chain, set up two-dimensional Markov chain.

Preferably, in above-mentioned inking device,

Described amending unit, when being more than or equal to described channel thresholding L specifically for voice traffic channel number current in systems in which, reduces the arrival rate of data user; When described voice traffic channel number is less than described channel thresholding L, reduce the service rate of data user.

Preferably, in above-mentioned inking device,

Described amending unit, also for by the arrival rate of data user by λ _ibe reduced to λ _iq _j, wherein, q _jbe one and be less than 1 and the function reduced along with the increase of j; And, by the service rate of data user by μ _ij () is reduced to μ _i(j) k _j, wherein, k _jbe one and be less than 1 and the function reduced along with the increase of j.

Preferably, in above-mentioned inking device,

Described system performance parameter comprises the probability of substitute of data service, the average queue length of data service, at least one in the blocking rate of the average delay of packet in queue and speech business calling.

As can be seen from the above, the collocation method of dynamic data Traffic Channel provided by the invention and device, achieve the optimization computation to dynamic PDCH allocation proportion innovatively, by dynamic PDCH allocation proportion is incorporated in Markov model, and calculated by numerical computations and system emulation and meeting under system channel performance constraints, make the dynamic PDCH of radio bearer efficiency optimization configure optimal value, thus the wireless channel load-carrying efficiency of system can be improved.And, the present embodiment considers that the punishment of wireless channel enters and time delay releasing mechanism, utilize existing network half rate ratio to enter the thresholding with time delay releasing mechanism as the punishment of introducing wireless channel, make dynamic PDCH allocation proportion result more can meet existing network statistics and existing network actual environment.The present embodiment can realize the automatic configuration of dynamic PDCH allocation proportion, can realize the intellectuality of district dynamic PDCH allocation proportion parameter and automation adjustment, alleviates workload that network optimization personnel directly configure and promotes the accuracy of configuration.

Accompanying drawing explanation

Fig. 1 is voice-and-data traffic channel assignments schematic diagram in prior art;

The collocation method that Fig. 2 is PDCH dynamic described in the embodiment of the present invention realize block diagram;

Fig. 3 is the schematic flow sheet of the collocation method of PDCH dynamic described in the embodiment of the present invention;

Fig. 4 be the collocation method of dynamic PDCH in the embodiment of the present invention realize block diagram;

Fig. 5 is the partial schematic diagram of the two-dimensional Markov chain before revising in the embodiment of the present invention;

Fig. 6 carries out to Fig. 5 the partial schematic diagram that data user accesses the two-dimensional Markov chain after punishment;

Fig. 7 is partial schematic diagram Fig. 5 being carried out to the two-dimensional Markov chain after data user's time delay whether;

Fig. 8 is the schematic diagram of the revised two-dimensional Markov chain of the embodiment of the present invention;

Fig. 9 is the structural representation of the inking device of dynamic PDCH in the embodiment of the present invention.

Embodiment

The collocation method of dynamic PDCH described in the embodiment of the present invention, target is turned to the maximum of radio bearer efficiency, with system performance requirements (such as, admissible audio call blocking rate, packet are substitute rate and Packet Delay) be constraints, according to configuration data and the half rate ratio of TCH real-time in system and fixed data Traffic Channel, automatically calculate the optimal value of the allocation proportion of dynamic PDCH, be applied to the adjustment of real-time existing network cell parameter, to promote radio bearer efficiency.

What Fig. 2 gave the collocation method of dynamic PDCH described in the embodiment of the present invention realizes block diagram, wherein left-half block representation is after dynamic channel ratio adjustment timer reaches predetermined value, the inking device of the dynamic PDCH of Automatically invoked right half part calculates, and according to the dynamic PDCH allocation ratio calculated, the relevant parameter of automatic configuration device.The inking device of dynamic PDCH is in computational process, according to existing network community real-time factor, obtain the parameter needed for modeling, and setting punishment access and time delay release rule, set up two-dimentional Markov model accordingly, the optimal value of dynamic PDCH allocation proportion is calculated, as the allocation ratio of dynamic PDCH channel under the different parameters of existing network cell equipment.

Because GPRS/EDGE introduces on original GSM network basis, limited channel resource is shared in data service and speech business, can do not distributed to GPRS/EDGE dynamically as required by the channel that speech business takies, this dynamic process can describe with the Markov process of a moving boundary.The present embodiment is from the channel allocation mechanism of equipment, set up a two-dimentional continuous time Markov chain, then under the different parameters of existing network cell equipment, the method solving system of linear equations is adopted to obtain the steady state solution of this Markov chain, some leading indicators of network performance are obtained finally by numerical computations, as: call blocking rate, data average transmission rate, channel utilization etc., select the dynamic PDCH allocation proportion with optimum radio bearer efficiency meeting predetermined network performance requirement.Please refer to Fig. 3, the collocation method of dynamic PDCH described in the embodiment of the present invention, can be applicable to a gprs system, the method specifically comprises the following steps:

Step 31, according to user behavior statistics in community, calculates voice user in described community, average arrival rate that data user enters system and by the average service rate of system service.

Step 32, for described community, according to average arrival rate and the average service rate of voice user, data user, the two-dimensional Markov chain that to set up with the data user's number in voice traffic channel number current in system, data user's waiting list be two-dimensional parameter.

Step 33, according to the magnitude relationship of described voice traffic channel number and channel thresholding L, revise arrival rate and the service rate of data user in described two-dimensional Markov chain, wherein, described channel thresholding L=N (1-δ), N is the total channel number of data and voice sharing in described community, and δ is the half rate ratio δ of voice channel in described community.

Here, in the present embodiment step 33, when described voice traffic channel number is more than or equal to described channel thresholding L, reduce the arrival rate of data user; When described voice traffic channel number is less than described channel thresholding L, reduce the service rate of data user.

Step 34, under different dynamic PDCH allocation proportions, calculate the steady state solution of described two-dimensional Markov chain, obtain the state transition probability matrix that two-dimensional Markov chain is corresponding, and based on described state transition probability matrix, computing system performance parameter, and judge whether the system performance parameter calculated meets predetermined system performance requirements, when meeting predetermined system performance requirements, calculate the load-carrying efficiency of the wireless channel under this dynamic PDCH allocation proportion.

Step 35, selects optimum wireless channel load-carrying efficiency, by the dynamic PDCH allocation proportion corresponding to the load-carrying efficiency of the wireless channel of this optimum, as the allocation ratio of the dynamic PDCH of gprs system.

The core of dynamic PDCH collocation method described in the present embodiment is to set up a model based on Markov chain improved, and carries out system emulation by solving of model.By setting Different Dynamic PDCH allocation proportion, the simulation result that comparison is different, draws the optimal value of dynamic PDCH allocation proportion under present channel environment.Specific algorithm schematic block diagram is as shown in Figure 4:

The output of the input of this algorithm, algorithm key technology and algorithm is described below respectively:

(1) algorithm input comprises:

1. the community to be adjusted TCH number of channel, the fixing PDCH number of channel;

2. cell system performance requirement, such as audio call loss probability, packet are substitute probability etc.

(2) the band punishment based on Markov chain enters the dynamic PDCH apportion model with time delay releasing mechanism

Because GPRS/EDGE introduces on original GSM network basis, limited channel resource is shared in data service and speech business, can do not distributed to GPRS/EDGE dynamically as required by the channel that speech business takies, this dynamic process can describe by the Markov process of a moving boundary.First from the channel allocation mechanism of equipment, set up a two-dimentional continuous time Markov chain, then the method solving system of linear equations is adopted to obtain the steady state solution of this Markov chain, some leading indicators of network performance are obtained finally by numerical computations, as: call blocking rate, data average transmission rate, channel utilization etc.And then judge whether these indexs meet predetermined system performance requirements, the load-carrying efficiency of wireless channel is calculated further when meeting, final selection has the dynamic PDCH allocation proportion of the highest radio bearer efficiency, as the allocation proportion of the dynamic PDCH of gprs system.

As can be seen from above step, the collocation method of the dynamic PDCH that the present embodiment provides, achieve the optimization computation to dynamic PDCH allocation proportion innovatively, by dynamic PDCH allocation proportion is incorporated in Markov model, and calculated by numerical computations and system emulation and meeting under system channel performance constraints, make the dynamic PDCH of radio bearer efficiency optimization configure optimal value, thus the wireless channel load-carrying efficiency of system can be improved.And, the present embodiment considers that the punishment of wireless channel enters and time delay releasing mechanism, utilize existing network half rate ratio to enter the thresholding with time delay releasing mechanism as the punishment of introducing wireless channel, make dynamic PDCH allocation proportion result more can meet existing network statistics and existing network actual environment.The present embodiment can realize the automatic configuration of dynamic PDCH allocation proportion, can realize the intellectuality of district dynamic PDCH allocation proportion parameter and automation adjustment, alleviates workload that network optimization personnel directly configure and promotes the accuracy of configuration.

Below explanation is specifically done to above-mentioned steps.

In the present embodiment, setting TCH number is C, FPDCH number is D.In above-mentioned steps 31, obtain the statistics of user behavior in community, such as, in the unit interval, have how many audio calls to enter system, in the unit interval, have how many data users to enter the data such as service time of system, audio call and data user.Speech business has priority, and under the prerequisite meeting speech business, data service may not be assigned to channel immediately, and is in queueing condition.Theoretical according to queueing theory, assuming that the arrival rate of speech business and data service all obeys Poisson (Poisson) distribution in system, and service time all obey quantum condition entropy, there is limiting value in the waiting list of packet, therefore utilize the statistical method of prior art, calculate the average arrival rate that voice user and data user enter system and be respectively λ _vand λ _d, be respectively μ by the average service rate of system service _vand μ _d.

In above-mentioned steps 32, for two-dimensional Markov chain is set up in described community.State P in this Markov chain _{i, j}in expression system, the current voice service number of channel is i, and the data user's number in data user's waiting list is the state probability of j.Suppose that the total channel number of data and voice sharing in described community is N, data user's number is M, for the state P in chain _{i, j}, then indicate that i data user enters into system (waiting in queue), M-i data user serves, and therefore data user's arrival rate can be described as: λ _i=(M-i) λ _d; The service rate of data user is then μ _i(j)=min (N-j, D+ η C, i) μ _d, i=0,1 ..., M, j=0,1 ..., N, voice traffic channel number current in j expression system here.For representing convenient, below by μ _ij () is referred to as μ _i.Speech business has priority relative to data service, for state P _{i, j}, the arrival rate of its speech business and service rate are then λ _vand μ _v.

Therefore, with the arrival rate λ of described data user _iwith service rate μ _i(j), as the transition probability of data user's number increase and decrease in two-dimensional Markov chain, and, with the average arrival rate of described voice user and average service rate, as the transition probability of voice traffic channel number increase and decrease in two-dimensional Markov chain, two-dimensional Markov chain (Fig. 5 illustrate only the part in chain) as shown in Figure 5 can be set up.Wherein, from P _{i-1, j}to P _{i, j}transition probability be λ _i-1, from P _{i, j}to P _{i+1, j}transition probability be λ _i, from P _{i+1, j}to P _{i, j}transition probability be μ _i+1, from P _{i, j}to P _{i-1, j}transition probability be μ _i; From P _{i, j-1}to P _{i, j}transition probability be λ _v, from P _{i, j}to P _{i, j+1}transition probability be λ _v, from P _{i, j+1}to P _{i, j}transition probability be μ _v, from P _{i, j}to P _{i, j-1}transition probability be μ _v.

The present embodiment further contemplates the access punishment of data user and time delay release, thus revises the arrival rate of data user in the described two-dimensional Markov chain shown in Fig. 5 and service rate in step 33, improves Markov chain.Wherein, when the current voice service number of channel is more than or equal to channel thresholding L in systems in which, reduce the arrival rate of data user; When the current voice service number of channel is less than described channel thresholding L in systems in which, reduce the service rate of data user.Below be explained respectively.

In the present embodiment, arrange a channel thresholding L=N (1-δ), here, N is the total channel number of data and voice sharing in described community, N=C+D, the half rate ratio δ of wherein TCH number is C, FPDCH number to be D, δ be voice channel in described community.When δ=0, system does not enable half rate, then think that system is completely not congested, then L=N; When δ=1, the whole half rate of system, then think that system is completely congested, then L=0.

In the Markov model that the present embodiment improves; the channel gate threshold value L started by arranging triggering voice congestion control introduces punishment and enters mechanism: when the voice channel number of access current in system meets or exceeds L; system is in congestion state; the number of channel that can be used for data service reduces; now start congestion control; adopt penalty mechanism to data arrival rate of customers, lower the data service probability of follow-up access, to ensure speech business when system congestion.And when the voice channel number of access current in system is less than L, system is in not congestion state, now then data user is adopted to the time delay releasing mechanism of channel, therefore reduce the service rate of data user, to improve the perception of data user when system is not congested.

Wherein, the process of punishing the arrival rate of data user is as follows:

I) data user sends session request (Data Session) by packet data protocol (PDP), and waiting system responds;

Ii) system mode confirms, if the current voice channel number of system is less than L, namely system is not congested, and now data service request ensures to transmit;

Iii) if system congestion, then dynamic assignment starts, and surely controls to punish parameter based on Congestion Level SPCC.

First M/M/N/N queuing model is adopted to carry out modeling, the representative of this model is input as M, service is the system of N number of information desk of M, in the present embodiment, and the expression data user number that M is concrete, N represents the total channel number of data and voice sharing, N=C+D, wherein TCH number is C, FPDCH number is D, if dynamically PDCH allocation proportion is η, then can show that in N, data service can use the interval of number of channel D ' for D≤D '≤D+ η C.

The present embodiment, in the M/M/N/N queuing model set up, improves the arrival rate under system congestion state, introduces penalty mechanism under system congestion to the arrival rate of data user.Concrete penalty mechanism can have multiple, its objective is the arrival rate making to reduce data user when system congestion.Below provide several penalty mechanism of the present embodiment:

1) such as, setting a demand function is

Here, p ₀be system not congested time controling parameters initial value, p _hcontroling parameters value when () is system congestion j, p _h(j)>=p ₀. be the apportionment ratio of the number of users under dynamic assignment, will the arrival process of data user under congestion state be determined.Setting q _j=(N-j) Δ q (j), wherein, the punishment that user is assigned to.Here the q calculated _jalways one is less than the number of 1, and is into the relation of negative correlation with j, namely reduces along with the increase of j.

When the current voice traffic channel taken is more than or equal to L in systems in which, system be in congested in, for ensureing that voice user accesses successfully, reduce data user arrival rate, to decrease the data service customers number of connecting system.Therefore, when i>=L, the arrival rate of the data user in Markov chain is revised, wherein, for state p _{i, j}, by the arrival rate of its data user by λ _ibe modified to λ _iq _j.

2) more such as, setting demand function is

Q _j=(N-j) Δ q (j), wherein, the punishment that user is assigned to.Here the q calculated _jalways one is less than the number of 1, and is into the relation of negative correlation with j, namely reduces along with the increase of j.Then, when i>=L, the arrival rate of the data user in Markov chain is revised, wherein, for state p _{i, j}, by the arrival rate of its data user by λ _ibe modified to λ _iq _j.

3) again such as, a constant Δ q is set, Δ q < 1/N;

Then q is calculated _j=(N-j) Δ q, the q calculated here _jalways one is less than the number of 1, and is into the relation of negative correlation with j, namely reduces along with the increase of j.Then, when i>=L, the arrival rate of the data user in Markov chain is revised, wherein, for state P _{i, j}, by the arrival rate of its data user by λ _ibe modified to λ _iq _j.

More than illustrate only for illustration of the specific implementation of punishment access, the present invention is not limited to above-mentioned situation.When the punishment access carrying out data user, such as, to P _{i, j}carry out data user when punishing access, only need to be multiplied by a weight being less than 1, this weight is the function reduced along with the increase of j.

Please refer to Fig. 6, show and the state transitions schematic diagram after punishment is accessed, wherein, from P to the data user of Fig. 5 _{i-1, j}to P _{i, j}transition probability be λ _i-1q _j, from P _{i, j}to P _{i+1, j}transition probability be λ _iq _j.Relative to Fig. 5, be all multiplied by weight q _j, i is the number being more than or equal to L here.

In the present embodiment, under system not congestion situation, time delay releasing mechanism is introduced to data user's service rate, namely time delay release process is adopted to the channel of data user, improve the service rate of system.Punishing in the model that enters, the service rate in model is that service completes and namely discharges channel, but in systems in practice, the release of channel often all can have certain time delay, and therefore the present embodiment considers delay strategy, the change of the service rate of improvement data user.Below provide several time delay releasing mechanism of the present embodiment:

1) such as, arranging demand function is m _h(j)>=m ₀,

M ₀be system not congested time controling parameters initial value, m _hcontroling parameters value when () is system congestion j.

K is the release rate of the number of users under dynamic assignment, and k will determine under not congestion state, the release delay of data user.Setting k _j=(N-j) Δ k (j), it is the burst size that user serves.Here the k calculated _jalways one is less than the number of 1, and is into the relation of negative correlation with j, namely reduces along with the increase of j.Then, when i < L, the service rate of the data user in Markov chain is revised, wherein, for state P _{i, j}, by the service rate of its data user by λ _ibe modified to μ _ik _j.

2) more such as, setting demand function is

K _j=(N-j) Δ k (j), wherein, it is the burst size that user serves.Here the k calculated _jalways one is less than the number of 1, and is into the relation of negative correlation with j, namely reduces along with the increase of j.Then, when i < L, the service rate of the data user in Markov chain is revised, wherein, for state p _{i, j}, by the service rate of its data user by μ _ibe modified to μ _ik _j.

3) again such as, a constant Δ k is set, Δ q < 1/N;

Then k is calculated _j=(N-j) Δ k, the k calculated here _jalways one is less than the number of 1, and is into the relation of negative correlation with j, namely reduces along with the increase of j.Then, when i < L, the service rate of the data user in Markov chain is revised, wherein, for state p _{i, j}, by the service rate of its data user by λ _ibe modified to μ _ik _j.

It should be noted that, the μ in Fig. 6 and hereinafter Fig. 7 _iall μ _ij writing a Chinese character in simplified form of (), the value of j is here the voice traffic channel number of the initial state of arrow.

Please refer to Fig. 7, show the state transitions schematic diagram after to data user's service rate time delay release of Fig. 5, wherein, from P _{i+1, j}to p _{i, j}transition probability be μ _i+1k _j, from P _{i, j}to P _{i-1, j}transition probability be μ _ik _j.Relative to Fig. 5, be all multiplied by weight k _j, i is the number of little L here.

Please refer to Fig. 8; give the two-dimensional Markov chain after the present embodiment introducing penalty mechanism and time delay releasing mechanism; wherein current in longitudinal axis expression system voice traffic channel number, transverse axis represent the data user's number in data user's waiting list; each ellipse circle represents a state, the transition probability between the Parametric Representation state between state on arrow.

After obtaining the Markov chain shown in Fig. 8, the steady state solution of described two-dimensional Markov chain under different dynamic PDCH allocation proportion η can be calculated, obtain the state transition probability matrix that two-dimensional Markov chain is corresponding, then based on described state transition probability matrix, judge whether to meet predetermined system performance parameter, when meeting predetermined system performance parameter, calculating the load-carrying efficiency of the wireless channel under this dynamic PDCH allocation proportion, being described as follows:

According to the own characteristic of two-dimensional Markov chain, for any state, all there is the balance entering this shape probability of state and leave between this shape probability of state, therefore can obtain following system of linear equations from Fig. 6:

Work as i=0, during j=0, (λ _v+ λ ₀) P _0,0=μ _vp _0,1+ μ ₁k ₀p _1,0

Work as i=0, during j=j, (λ _v+ λ ₀+ μ _v) P _{0, j}=μ _vp _{0, j+1}+ μ ₁k _jp _{1, j}+ λ _vp _{0, j-1}

Work as i=0, during j=L, (λ _v+ λ ₀q _l+ μ _v) P _{0, L}=μ _vp _{0, L+1}+ μ ₁p _{1, L}+ λ _vp _{0, L-1}

Work as i=0, during j=N, (λ ₀q _n+ μ _v) P _{0, N}=λ _vp _{0, N-1}

Work as i=i, during j=0, (λ _v+ λ _i+ μ _ik ₀) P _{i, 0}=μ _vp _{i, 1}+ μ _i+1k ₀p _i+1,0+ λ _i-1p _i-1,0

When i ≠ 0, during j < L, (λ _v+ λ _i+ μ _v+ μ _ik _j) P _{i, j}=μ _vp _{i, j+1}+ μ _i+1k _jp _{i+1, j}+ λ _vp _{i, j-1}+ λ _i-1p _{i-1, j}

When i ≠ 0, during j=L, (λ _v+ λ _iq _l+ μ _v+ μ _i) P _{i, L}=μ _vp _{i, L+1}+ μ _i+1p _{i+1, L}+ λ _vp _{i, L-1}+ λ _i-1q _lp _{i-1, L}

When i ≠ 0, during j=N, (λ _iq _n+ μ _v) P _{i, N}=λ _vp _{i, N-1}+ λ _i-1q _np _{i-1, N}

Here, set p=[ p ₀ , p ₁ ..., p _m ], p _i =[P _{i, 0}, P _{i, 1}..., P _{i, N}] ^t, pbe the state transition probability matrix of two-dimensional Markov chain, there is normalization characteristic.

Exist pk=0, here, matrix K is R × R battle array, R=(M+1) N, and

$K=\left[\begin{array}{cccccccc}{A}_0& V_0& 0& \&CenterDot;\&CenterDot;\&CenterDot;& & & & 0\\ U_1& A_1& V_1& 0& \&CenterDot;\&CenterDot;\&CenterDot;& & & 0\\ 0& U_2& A_2& V_2& 0& \&CenterDot;\&CenterDot;\&CenterDot;& & \\ 0& & & & & & & V_{M-1}\\ 0& \&CenterDot;\&CenterDot;\&CenterDot;& & & & 0& U_M& A_M\end{array}\right]$

Wherein,

$A_{i=}\left[\begin{array}{cccccc}{\mathrm{\&lambda;}}_v+{\mathrm{\&lambda;}}_i+{\mathrm{\&mu;}}_1{k}_0& -{\mathrm{\&lambda;}}_v& 0& 0& 0& 0\\ -{\mathrm{\&mu;}}_v& {\mathrm{\&lambda;}}_v+{\mathrm{\&lambda;}}_i+{\mathrm{\&mu;}}_v+{\mathrm{\&mu;}}_2{k}_1& & & & \\ 0& -{\mathrm{\&mu;}}_v& & & & \\ & 0& & -{\mathrm{\&lambda;}}_v& & \\ & & & {\mathrm{\&lambda;}}_v+{\mathrm{\&lambda;}}_i{q}_L+{\mathrm{\&mu;}}_v+{\mathrm{\&mu;}}_L& & \\ & & & -{\mathrm{\&mu;}}_v& & \\ & & & & & -{\mathrm{\&lambda;}}_v\\ 0& 0& \&CenterDot;\&CenterDot;\&CenterDot;& & & {\mathrm{\&lambda;}}_i{q}_N+{\mathrm{\&mu;}}_v\end{array}\right]$

$U_i=\left[\begin{array}{cccccc}-{\mathrm{\&mu;}}_1{k}_0& 0& 0& \&CenterDot;\&CenterDot;\&CenterDot;& & 0\\ 0& -{\mathrm{\&mu;}}_2{k}_1& & & & \&CenterDot;\&CenterDot;\&CenterDot;\\ \&CenterDot;\&CenterDot;\&CenterDot;& 0& -{\mathrm{\&mu;}}_3{k}_2& & & \\ & & & \&CenterDot;\&CenterDot;\&CenterDot;& & \\ & & & & -{\mathrm{\&mu;}}_{M-1}& 0\\ 0& 0& \&CenterDot;\&CenterDot;\&CenterDot;& & 0& 0\end{array}\right]$

$V_i=\left[\begin{array}{cccccc}-{\mathrm{\&lambda;}}_0& 0& 0& \&CenterDot;\&CenterDot;\&CenterDot;& & 0\\ 0& -{\mathrm{\&lambda;}}_1& & & & \&CenterDot;\&CenterDot;\&CenterDot;\\ \&CenterDot;\&CenterDot;\&CenterDot;& 0& \&CenterDot;\&CenterDot;\&CenterDot;& & & \\ & & & -{\mathrm{\&lambda;}}_{i-1}{q}_L& & \\ & & & & \&CenterDot;\&CenterDot;\&CenterDot;& 0\\ 0& 0& \&CenterDot;\&CenterDot;\&CenterDot;& & 0& -{\mathrm{\&lambda;}}_{M-1}{q}_N\end{array}\right]$

Therefore, under the dynamic PDCH allocation proportion that certain is given, above-mentioned two-dimensional Markov chain is solved, the state transition probability matrix of two-dimensional Markov chain can be obtained p.Based on this matrix, system performance parameter can be calculated, the probability of substitute of such as data service, the average queue length of data service, the average delay of packet in queue and the blocking rate etc. of speech business calling.Such as,

The probability of substitute of data service is

The average queue length of data service is

The average delay of data service is

The blocking rate of speech business calling is

Here, i=0,1 ..., M, j=0,1 ..., N.

Then, judge whether the above-mentioned system performance parameter calculated meets predetermined system performance requirements, if met, then calculate the load-carrying efficiency of the wireless channel under current dynamic PDCH allocation proportion further, specific formula for calculation is as follows:

Speech business TCH channel utilization is

FPDCH channel utilization is fixed in data service

Then total wireless channel load-carrying efficiency is:

Here, C represents the TCH number of channel, and D represents the FPDCH number of channel.

Finally, in the wireless channel carrying ratio obtained under Different Dynamic PDCH allocation proportion, choose an optimal value, by dynamic PDCH allocation proportion corresponding for this optimal value, as the allocation proportion of the dynamic PDCH of gprs system.

Finally, based on the collocation method of above-mentioned dynamic PDCH, the present embodiment also accordingly provides a kind of inking device in order to realize above-mentioned collocation method, and as shown in Figure 9, the inking device of the dynamic PDCH that the present embodiment provides, comprising:

Computing unit, for according to user behavior statistics in community, calculates voice user in described community, average arrival rate that data user enters system and by the average service rate of system service;

Modeling unit, for for described community, according to average arrival rate and the average service rate of voice user, data user, the two-dimensional Markov chain that to set up with the data user's number in voice traffic channel number current in system, data user's waiting list be two-dimensional parameter;

Amending unit, for the magnitude relationship according to voice traffic channel number current in system and channel thresholding L, revise arrival rate and the service rate of data user in described two-dimensional Markov chain, wherein said channel thresholding L=N (1-δ), N is the total channel number of data and voice sharing in described community, and δ is the half rate ratio of voice channel in described community;

Simulation unit, for under different dynamic PDCH allocation proportions, calculate the steady state solution of described two-dimensional Markov chain, obtain the state transition probability matrix that two-dimensional Markov chain is corresponding, and based on described state transition probability matrix, computing system performance parameter, and judge whether the system performance parameter calculated meets predetermined system performance requirements, when meeting described system performance requirements, calculate the load-carrying efficiency of the wireless channel under this dynamic PDCH allocation proportion;

Selected cell, for selecting optimum wireless channel load-carrying efficiency, by the dynamic PDCH allocation proportion corresponding to the load-carrying efficiency of the wireless channel of this optimum, as the allocation ratio of dynamic PDCH.

Preferably, described modeling unit, specifically for:

For each the state P in two-dimensional Markov chain _{i, j}, the arrival rate λ of calculated data user _i=(M-i) λ _dwith service rate μ _i(j)=min (N-j, D+ η C, i) μ _d, wherein, i=0,1 ..., M, represents the data user's number in data user's waiting list; J=0,1 ..., N, voice traffic channel number current in expression system, N represents the total channel number of data and voice sharing in described community, and M represents the data user's sum activated in described community, and C represents the TCH number of channel, D represents the FPDCH number of channel, and η represents dynamic PDCH allocation proportion, λ _dand μ _drespectively represent data user enter system average arrival rate and by the average service rate of system service;

With the arrival rate λ of described data user _iwith service rate μ _i(j), as the transition probability of data user's number increase and decrease in two-dimensional Markov chain, and, with the average arrival rate of described voice user and average service rate, as the transition probability of voice traffic channel number increase and decrease in two-dimensional Markov chain, set up two-dimensional Markov chain.

Preferably, described amending unit, when being more than or equal to described channel thresholding L specifically for voice traffic channel number current in systems in which, reduces the arrival rate of data user; When described voice traffic channel number is less than described channel thresholding L, reduce the service rate of data user.

Preferably, described amending unit, also for by the arrival rate of data user by λ _ibe reduced to λ _iq _j, wherein, q _jbe one and be less than 1 and the function reduced along with the increase of j; And, by the service rate of data user by μ _ij () is reduced to μ _i(j) k _j, wherein, k _jbe one and be less than 1 and the function reduced along with the increase of j.

Preferably, described system performance parameter comprises the probability of substitute of data service, the average queue length of data service, at least one in the blocking rate of the average delay of packet in queue and speech business calling.

The above is only embodiments of the present invention; it should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (8)

1. a collocation method of dynamic PDCH, is characterized in that, comprising:

According to user behavior statistics in community, calculate voice user in described community, average arrival rate that data user enters system and by the average service rate of system service;

For described community, according to average arrival rate and the average service rate of voice user, data user, the two-dimensional Markov chain that to set up with the data user's number in voice traffic channel number current in system, data user's waiting list be two-dimensional parameter;

According to the magnitude relationship of voice traffic channel number current in system and channel thresholding L, revise arrival rate and the service rate of data user in described two-dimensional Markov chain, wherein said channel thresholding L=N (1-δ), N is the total channel number of data and voice sharing in described community, and δ is the half rate ratio of voice channel in described community;

Under different dynamic PDCH allocation proportions, calculate the steady state solution of described two-dimensional Markov chain, obtain the state transition probability matrix that two-dimensional Markov chain is corresponding, and based on described state transition probability matrix, computing system performance parameter, and judge whether the system performance parameter calculated meets predetermined system performance requirements, when meeting described system performance requirements, calculates the load-carrying efficiency of the wireless channel under this dynamic PDCH allocation proportion;

Select optimum wireless channel load-carrying efficiency, by the dynamic PDCH allocation proportion corresponding to the load-carrying efficiency of the wireless channel of this optimum, as the allocation ratio of dynamic PDCH;

Wherein, the two-dimensional Markov chain that described foundation is two-dimensional parameter with the data user's number in voice traffic channel number current in system, data user's waiting list, comprising:

For each the state P in two-dimensional Markov chain _i,j, the arrival rate λ of calculated data user _i=(M-i) λ _dwith service rate μ _i(j)=min (N-j, D+ η C, i) μ _d, wherein, i=0,1 ..., M, represents the data user's number in data user's waiting list; J=0,1 ..., N, voice traffic channel number current in expression system, N represents the total channel number of data and voice sharing in described community, and M represents the data user's sum activated in described community, and C represents the TCH number of channel, D represents the FPDCH number of channel, and η represents dynamic PDCH allocation proportion, λ _dand μ _drespectively represent data user enter system average arrival rate and by the average service rate of system service;

With the arrival rate λ of described data user _iwith service rate μ _i(j), as the transition probability of data user's number increase and decrease in two-dimensional Markov chain, and, with the average arrival rate of described voice user and average service rate, as the transition probability of voice traffic channel number increase and decrease in two-dimensional Markov chain, set up two-dimensional Markov chain.

2. collocation method as claimed in claim 1, is characterized in that,

The described magnitude relationship according to described voice traffic channel number and channel thresholding L, revise arrival rate and the service rate of data user in described two-dimensional Markov chain, comprising:

When voice traffic channel number current is in systems in which more than or equal to described channel thresholding L, reduce the arrival rate of data user; When described voice traffic channel number is less than described channel thresholding L, reduce the service rate of data user.

3. collocation method as claimed in claim 2, is characterized in that,

The arrival rate of described reduction data user is specially: by the arrival rate of data user by λ _ibe reduced to λ _iq _j, wherein, q _jbe one and be less than 1 and the function reduced along with the increase of j;

The service rate of described reduction data user is specially: by the service rate of data user by μ _ij () is reduced to μ _i(j) k _j, wherein, k _jbe one and be less than 1 and the function reduced along with the increase of j.

4. collocation method as claimed in claim 1, is characterized in that,

Described system performance parameter comprises the probability of substitute of data service, the average queue length of data service, at least one in the blocking rate of the average delay of packet in queue and speech business calling.

5. an inking device of dynamic PDCH, is characterized in that, comprising:

Computing unit, for according to user behavior statistics in community, calculates voice user in described community, average arrival rate that data user enters system and by the average service rate of system service;

Modeling unit, for for described community, according to average arrival rate and the average service rate of voice user, data user, the two-dimensional Markov chain that to set up with the data user's number in voice traffic channel number current in system, data user's waiting list be two-dimensional parameter;

Amending unit, for the magnitude relationship according to voice traffic channel number current in system and channel thresholding L, revise arrival rate and the service rate of data user in described two-dimensional Markov chain, wherein said channel thresholding L=N (1-δ), N is the total channel number of data and voice sharing in described community, and δ is the half rate ratio of voice channel in described community;

Simulation unit, for under different dynamic PDCH allocation proportions, calculate the steady state solution of described two-dimensional Markov chain, obtain the state transition probability matrix that two-dimensional Markov chain is corresponding, and based on described state transition probability matrix, computing system performance parameter, and judge whether the system performance parameter calculated meets predetermined system performance requirements, when meeting described system performance requirements, calculate the load-carrying efficiency of the wireless channel under this dynamic PDCH allocation proportion;

Selected cell, for selecting optimum wireless channel load-carrying efficiency, by the dynamic PDCH allocation proportion corresponding to the load-carrying efficiency of the wireless channel of this optimum, as the allocation ratio of dynamic PDCH; Wherein, described modeling unit, specifically for:

For each the state P in two-dimensional Markov chain _i,j, the arrival rate λ of calculated data user _i=(M-i) λ _dwith service rate μ _i(j)=min (N-j, D+ η C, i) μ _d, wherein, i=0,1 ..., M, represents the data user's number in data user's waiting list; J=0,1 ..., N, voice traffic channel number current in expression system, N represents the total channel number of data and voice sharing in described community, and M represents the data user's sum activated in described community, and C represents the TCH number of channel, D represents the FPDCH number of channel, and η represents dynamic PDCH allocation proportion, λ _dand μ _drespectively represent data user enter system average arrival rate and by the average service rate of system service;

With the arrival rate λ of described data user _iwith service rate μ _i(j), as the transition probability of data user's number increase and decrease in two-dimensional Markov chain, and, with the average arrival rate of described voice user and average service rate, as the transition probability of voice traffic channel number increase and decrease in two-dimensional Markov chain, set up two-dimensional Markov chain.

6. inking device as claimed in claim 5, is characterized in that,

Described amending unit, when being more than or equal to described channel thresholding L specifically for voice traffic channel number current in systems in which, reduces the arrival rate of data user; When described voice traffic channel number is less than described channel thresholding L, reduce the service rate of data user.

7. inking device as claimed in claim 6, is characterized in that,

Described amending unit, also for by the arrival rate of data user by λ _ibe reduced to λ _iq _j, wherein, q _jbe one and be less than 1 and the function reduced along with the increase of j; And, by the service rate of data user by μ _ij () is reduced to μ _i(j) k _j, wherein, k _jbe one and be less than 1 and the function reduced along with the increase of j.

8. inking device as claimed in claim 5, is characterized in that,

Described system performance parameter comprises the probability of substitute of data service, the average queue length of data service, at least one in the blocking rate of the average delay of packet in queue and speech business calling.