CN103945477B - Heterogeneous network selection system and method based on business type weight differentiation - Google Patents

Abstract

The invention discloses a heterogeneous network selection system and method based on business type weight differentiation, relates to the field of heterogeneous network selection, and aims to solve the problems that the lowest QoS (quality of service) threshold values of business requirements are not different, and network utilization rate is reduced and the quality of service is weakened as loads between networks are unbalanced. The heterogeneous network selection system comprises a network discovery and static information collection module (1), a candidate network QoS parameter value acquisition module (2) and a candidate network evaluation and selection module (3). The method particularly includes the steps: firstly, finding adjacent RANs (radio access networks), collecting static information of the adjacent RANs and storing the static information in a database by an MIHF (media independent handover function); secondly, bringing the detected adjacent RANs positioned on a communication threshold value into a candidate network rank, and collecting QoS parameter values of the candidate network rank; thirdly, selecting a candidate network with the largest ideal approximation value to serve as a target network.

Description

It is a kind of that system and method is selected based on the heterogeneous network that type of service weights are distinguished

Technical field

The present invention relates to a kind of select system and method based on the heterogeneous network that type of service weights are distinguished.

Background technology

As the globalization and the fast development of Internet of telecommunication industry, communication system emerge in an endless stream, form various each The separate network of sample, such as cell mobile communication systems (2G, 3G), short distance access system (WLAN, bluetooth, ultra broadband UWB systems), WiMAX, satellite communication, broadcast system (such as DAB, DVB), stratosphere etc..The network situation of these networks is mutually only Vertical, every kind of network has its respective advantage, and the network communicating system of such as 2G and 3G is based on voice communication, and supports that data are led to Letter；And the network such as WLAN, WiMAX is then based on data service.With the gradually atrophy of voice communication service, wireless data is passed The development of defeated business, the personalized multifarious business demand of user constantly strengthen, and the network mould of this single network list application Formula exposes its intrinsic drawback in terms of network share：Different communication protocol causes the shared realization of wireless network difficult, unfavorable In data service sound cross network transmission.

In order to, while protecting legacy network to invest, accommodate the mobile communication subscriber for increasing sharply, and provide high-quality shifting Dynamic operation service in multimedia, overlapped but different types of various cordless communication networks are necessarily fused together collaborative work, Form unified, based on IP agreement heterogeneous wireless networks system.Mobile terminal also will towards integrate various access technologies, The multimode intelligent terminal direction of intelligent selection access network is developed.

Using internetwork complementary characteristic, the seamless fusion of various wireless networks is realized, personalized service is provided the user, is The inevitable direction of future communications network Development.Under heterogeneous wireless networks environment, how mobile terminal selects most effective, most suitable Access network, to meet the service quality (QualityofService, QoS) of business, study hotspot always in recent years.

The scheme of heterogeneous network converged, for a user, user can be according to factors such as business demand and network states Select at any time to get involved in most suitable network, so as to meet oneself flexible and changeable individual business demand.And for network For operation, using heterogeneous network converged scheme, operator does not need substantial amounts of fund input, and laying again is set up network and set Apply, it is not necessary to the very long construction period, it is only necessary to add on the basis of oneself original system structure a small amount of network it is real and Corresponding management function can both meet the ever-increasing traffic demands of user, and improve the market competitiveness of system.

When traditional method is to meet QoS and carry out network as target and select, the often weight of each qos parameter of mean allocation Value, will cause no matter the minimum QoS threshold of which kind of business demand does not have difference, so initiating do not have spy when network is selected in terminal Fixed network preference, but the network for tending to select QoS optimum carries out service bearer, and all terminals will be caused all to access most Excellent network, so as to cause network congestion, and remaining network idle, load imbalance between all-network is caused, while optimal network Congestion will reduce network utilization, so as to weaken service quality.

Difference due to not accounting for QoS demand between different kinds of business, so no matter which kind of business, always selects QoS optimum network insertion, so as to cannot rationally, effectively balance the load of each candidate network.

The content of the invention

The invention aims to the minimum QoS threshold for solving business demand does not have difference, load imbalance between network Reduce network utilization, weaken service quality problem and propose it is a kind of based on type of service weights distinguish heterogeneous network choosing Select system and method.

Above-mentioned goal of the invention is achieved through the following technical solutions：

It is a kind of to select system to specifically include based on the heterogeneous network that type of service weights are distinguished：Network finding and static information Collection module, candidate network qos parameter numerical value acquisition module and candidate network evaluation and selecting module；

Described Network finding and static information collection module find adjacent R AN by MIHF and collect adjacent R AN Static information；

Described candidate network qos parameter numerical value acquisition module detects Network finding and static information collection module successively The RSS of neighboring static information RAN is collected, the RSS of neighboring static information RAN after detection is included into candidate network ranks, is passed through MIHF collects the QoS parameter value of current candidate network ranks；

Described candidate network evaluation and selecting module are using working as that candidate network qos parameter numerical value acquisition module is obtained The QoS parameter value of front candidate network ranks calculates the ideal of candidate network and approaches the maximum candidate network of value as target network Network.

What a kind of heterogeneous network system of selection distinguished based on type of service weights was specifically followed the steps below：

Step one, the MIHF configured using the whole network in Network finding and static information collection module are determined in mobile terminal When plan is to communicate in current radio access network RAN, finds adjacent R AN and collect the static information of adjacent R AN, and phase will be collected The static information of adjacent RAN is stored in the local database；

Step 2, the wireless static that mobile terminal is detected current RAN using candidate network qos parameter numerical value acquisition module State information signal intensity RSS drops to when needing switching threshold, then mobile terminal initiates handover request to objective network, is switched to The wireless static information of current RAN in local data base, detects the RSS of the wireless static information of RAN successively, will detect and be in Adjacent R AN on communication threshold includes candidate network ranks, collects the QoS parameter value of candidate network ranks by MIHF；

Step 3, by the main business classification of current mobile terminal decision-making, according to the QoS parameter value profit of candidate network ranks The ideal that candidate network is obtained with AHP algorithms calculating qos parameter weight vector and evaluation algorithms calculating qos parameter weight vector is forced Close values, the judgement matrix after being configured to weight using the QoS parameter value of current candidate network ranks select ideal to approach value most Big candidate network, will select ideal to approach the maximum candidate network of value as objective network；Wherein, four kinds of types of service include Streaming Media class, interactive class, background class and conversation class；A kind of heterogeneous network selection distinguished based on type of service weights is completed Method.

Invention effect

The present invention arranges the QoS ginsengs per class business using analytic hierarchy process (AHP) (AnalyticHierarchyProcess, AHP) Number weights, then with ideal value approximation sort algorithm Based on (TechniqueforOrderPreferencebySimilaritytoIdeal Solution, TOPSIS), devise A kind of differentiation Weights-selected Algorithm of service-oriented classification carries out heterogeneous network selection.As traditional network selecting method is more with average Weights method carries out QoS of survice parameter setting, the QoS demand difference between not consideration business, concentrates and selects QoS optimal networks to enter Row service bearer, so as to cannot rationally, effectively balance the load of various candidate networks, so as to cause network congestion and resource Waste.And the present invention utilizes analytic hierarchy process (AHP) reasonable analysis, is provided with the QoS parameter value weight of the four class business that 3GPP is defined, Select different networks to carry out service bearer according to the QoS demand of different service types, comprehensively utilize the excellent of different access networks Gesture characteristic, realizes the comprehensive utilization of whole network, reaches the effect of resource optimization.

Instant invention overcomes conversation class and stream class service are accessed UMTS network by traditional average weight method in a large number, and will hand over The UMTS network long-term blockage changed class and background class traffic more multiple access wlan network and cause, reduce terminal access UMTS into The problem of power.The present invention is capable of achieving stream class and to exchange class business and more be switched to this relative High Data Rates of WiMAX, low In the network of the bit error rate, while QoS of survice is ensured, it is for difference of the different classes of business in QoS demand, right Four class business of 3GPP definition have carried out the analysis of qos parameter relative importance, set the qos parameter weights of every class business, close Reason selection target network is while the load of the multiple candidate networks of active balance.Competition has been embodied in terms of effectively utilizes Internet resources Power.

Present invention improves average traffic classification weights carry out offered load inequality caused by network is selected in traditional method Weighing apparatus problem, has carried out QoS weights classification with reference to AHP algorithms for 4 kinds of classs of service defined in 3GPP, for different business Classification calculates the QoS weight vectors needed for which, then carry out when network is selected can reasonable access network, and unlike in traditional method Access optimal network is all concentrated, but business reasonable layout is rationally effectively utilized in multiple networks by net according to type of service Network resource, such as balance network load, Fig. 4 and Fig. 5 can illustrate that the network in the present invention selects access result more rationally balanced, three Individual network effectively carries different service types, and concentrates in traditional method and access fast network 1 and cause network 3 idle.

Description of the drawings

Fig. 1 is a kind of heterogeneous network selection system work(distinguished based on type of service weights that specific embodiment two is proposed Can module map；Wherein, (1) is Network finding and static information collection module, and (2) are gathered for candidate network qos parameter numerical value Module, (3) are candidate network evaluation and selecting module；

Fig. 2 is a kind of heterogeneous network system of selection stream distinguished based on type of service weights that specific embodiment three is proposed Cheng Tu；

Fig. 3 is that the selection ideal that specific embodiment four is proposed approaches the maximum candidate network of value as objective network flow process Figure；

Fig. 4 is the average QoS weights net of justice network selection result figure that embodiment is proposed；Wherein, abscissa is four kinds of service class Type 1 be conversation class, 2 be Streaming Media class, 3 be interactive class and 4 be background class；Vertical coordinate is the probability distribution for selecting certain network； N1 is (UMTS) candidate network, and N2 is (WiMAX) candidate network, and N3 is (WLAN) candidate network；

Fig. 5 is the differentiated service QoS weights net of justice network selection result schematic diagrams that embodiment is proposed；Wherein, abscissa is four Kind of type of service 1 be conversation class, 2 be Streaming Media class, 3 be interactive class and 4 be background class；Vertical coordinate be select certain network it is general Rate is distributed；N1 is (UMTS) candidate network, and N2 is (WiMAX) candidate network, and N3 is (WLAN) candidate network；

Fig. 6 is that the present invention that embodiment is proposed selects number of times Statistical Comparison figure to network with average weight algorithm；Wherein, it is horizontal Coordinate N1 is (UMTS) candidate network, and N2 is (WiMAX) candidate network, and N3 is (WLAN) candidate network；Vertical coordinate is network quilt Select number of times.

Specific embodiment

Specific embodiment one：The a kind of of present embodiment selects system based on the heterogeneous network that type of service weights are distinguished Specifically include：

Network finding and static information collection module 1, candidate network qos parameter numerical value acquisition module 2 and candidate network are commented Valency and selecting module 3；

Described Network finding and static information collection module 1 find adjacent R AN by MIHF and collect adjacent R AN Static information；

Described candidate network qos parameter numerical value acquisition module 2 detects that Network finding and static information collect mould successively Block 1 collects the RSS of neighboring static information RAN, and the RSS of neighboring static information RAN after detection is included candidate network ranks, leads to Cross the QoS parameter value that MIHF collects current candidate network ranks；

What described candidate network evaluation and selecting module were obtained using candidate network qos parameter numerical value acquisition module 2 The QoS parameter value of current candidate network ranks calculates the ideal of candidate network and approaches the maximum candidate network of value as target network Network.

Present embodiment effect：

Present embodiment is arranged per class business using analytic hierarchy process (AHP) (AnalyticHierarchyProcess, AHP) Qos parameter weights, then with ideal value approximation sort algorithm Based on (TechniqueforOrderPreferencebySimilarityto IdealSolution, TOPSIS), devise A kind of differentiation Weights-selected Algorithm of service-oriented classification carries out heterogeneous network selection.As traditional network selecting method is more with average Weights method carries out QoS of survice parameter setting, the QoS demand difference between not consideration business, concentrates and selects QoS optimal networks to enter Row service bearer, so as to cannot rationally, effectively balance the load of various candidate networks, so as to cause network congestion and resource Waste.And present embodiment is using analytic hierarchy process (AHP) reasonable analysis, the QoS parameter value power of the four class business for being provided with 3GPP definition Weight, selects different networks to carry out service bearer according to the QoS demand of different service types, comprehensively utilizes different access networks Advantages characteristic, realizes the comprehensive utilization of whole network, reaches the effect of resource optimization.

Present embodiment overcomes traditional average weight method and conversation class and stream class service is accessed UMTS network in a large number, and The UMTS network long-term blockage that class and background class traffic more multiple access wlan network will be exchanged and caused, reduces terminal and accesses UMTS Success rate problem.Present embodiment is capable of achieving stream class and to exchange class business to be more switched to WiMAX this relatively high In data transfer rate, the network of low error rate, while QoS of survice is ensured, for different classes of business in QoS demand Four class business of 3GPP definition have been carried out the analysis of qos parameter relative importance, have set the qos parameter of every class business by difference Weights, reasonable selection objective network is while the load of the multiple candidate networks of active balance.Embody in terms of effectively utilizes Internet resources Competitiveness is gone out.

During present embodiment improves traditional method, average traffic classification weights carry out offered load caused by network is selected Unbalanced problem, has carried out QoS weights classification with reference to AHP algorithms for 4 kinds of classs of service defined in 3GPP, for different Class of service calculates the QoS weight vectors needed for which, then carry out when network is selected can reasonable access network, and unlike tradition side Access optimal network is concentrated in method all, but according to type of service by business reasonable layout in multiple networks, it is rationally effectively sharp With Internet resources, such as balance network load, Fig. 4 and Fig. 5, can illustrate that the network in present embodiment selects access result more reasonable Equilibrium, three networks effectively carry different service types, and concentrate in traditional method and access fast network 1 and cause network 3 It is idle.

Specific embodiment two：Present embodiment from unlike specific embodiment one：Described Network finding and It is current radio access network in mobile terminal decision-making that static information collection module 1 is used for the MIHF of the whole network configuration In (RadioAccess Network, RAN) during communication, find adjacent R AN and collect the static information of adjacent R AN, and will collect The static information of adjacent R AN is stored in the local database；Wherein, the switching mode of mobile terminal decision-making refers to that whole network is selected And final objective network is initiated and determined by mobile terminal to handover decisions, it is the leading behavior of terminal, is not to be determined by network Fixed；

Described candidate network qos parameter numerical value acquisition module 2 for detecting the wireless static of current RAN when mobile terminal The state information signal intensity signal intensity (RadioSignal that terminal antenna is received i.e. in current radio access network Strength, RSS) drop to when needing switching threshold, phase is detected successively according to the static information of adjacent R AN in local data base Adjacent R AN on communication threshold is included candidate network ranks, then collects current by MIHF by the RSS of adjacent RAN The QoS parameter value of candidate network ranks；

Described candidate network evaluation and selecting module 3 by the main business classification of current mobile terminal decision-making, according to The QoS parameter value of current candidate network ranks calculates qos parameter weight vector using AHP algorithms, and Utilization assessment algorithm is calculated Qos parameter weight vector obtains the ideal of candidate network and approaches value, is constructed using the QoS parameter value of current candidate network ranks Select ideal to approach the maximum candidate network of value into the judgement matrix after weighting, ideal will be selected to approach the maximum candidate network of value As objective network such as Fig. 1.Other steps and parameter are identical with specific embodiment one.

Specific embodiment three：Present embodiment from unlike specific embodiment one or two：It is a kind of to be based on service class The heterogeneous network system of selection that type weights are distinguished comprises the steps：

Step one, using in Network finding and static information collection module 1 the whole network configuration MIHF in mobile terminal Decision-making is in current radio access network (RadioAccessNetwork, RAN), during communication, discovery adjacent R AN simultaneously collects adjacent The static information of RAN, and the static information for collecting adjacent R AN is stored in the local database；Wherein MIHF is hypothesis the whole network Configured the Media Independent Handover function of IEEE802.21 (MediaIndependentHandover Function, MIHF)；

Step 2, the wireless static that mobile terminal is detected current RAN using candidate network qos parameter numerical value acquisition module State information signal intensity (RadioSignalStrength, RSS) drops to when needing switching threshold, then mobile terminal is to target Network initiates handover request, is switched to the wireless static information of current RAN in local data base, detects the wireless static of RAN successively The RSS of state information, will detect that adjacent R AN on communication threshold includes candidate network ranks, is collected by MIHF and is worked as The QoS parameter value of front candidate network ranks；

Step 3, by the main business classification of current mobile terminal decision-making, according to the QoS parameter value profit of candidate network ranks The ideal that candidate network is obtained with AHP algorithms calculating qos parameter weight vector and evaluation algorithms calculating qos parameter weight vector is forced Close values, the judgement matrix after being configured to weight using the QoS parameter value of current candidate network ranks select ideal to approach value most Big candidate network, will select ideal to approach the maximum candidate network of value as objective network；Wherein, AHP algorithms calculate four kinds The QoS weight vectors of type of service, the QoS weight vectors of calculate four kinds of types of service are stored in data base；Four kinds Type of service includes Streaming Media class, interactive class, background class and conversation class；Complete it is a kind of based on type of service weights distinguish Heterogeneous network system of selection such as Fig. 2.Other steps and parameter are identical with specific embodiment one or two.

Specific embodiment four：Unlike one of present embodiment and specific embodiment one to three：Will in step 3 The main business classification of current mobile terminal decision-making, calculates QoS using AHP algorithms according to the QoS parameter value of candidate network ranks Parameter weight vector and evaluation algorithms calculating qos parameter weight vector obtain the ideal of candidate network and approach value, using current Judgement matrix after the QoS parameter value of candidate network ranks is configured to weight selects ideal to approach the maximum candidate network of value, will Ideal is selected to approach the maximum candidate network of value as objective network；Wherein, AHP algorithms calculate the QoS power of four kinds of types of service Value vector, the QoS weight vectors of calculate four kinds of types of service are present in data base as Fig. 3 detailed processes are as follows：

(1) assume that the candidate network number that mobile terminal determines is m, using the candidate network QoS parameter value construction for obtaining Judgement matrix X'；

Wherein, in formula (1), each element is the raw value of qos parameter in candidate network, i-th candidate network N_iInstitute There is QoS parameter value represent d with the row vector of matrix X'_i'dj_i'plr_i'gb_i'mb_i' judgement matrix X' normalization is returned Judgement matrix X after one change；

D：Candidate network time delay, DJ：Candidate network delay variation, PLR：Candidate network packet loss, GB：Candidate network is ensured Bit rate, MB：Candidate network Maximum Bit Rate, i be i-th network, 1≤i≤m,

r_ijCorresponding QoS parameter value corresponding to i-th network, wherein, m is candidate network number, and j is candidate network parameter That is the columns of X matrix；x_ijFor the element of the i-th row of X matrix jth row, j-th QoS parameter value normalized value of as i-th network；

2) it is standardized to adjudicating the element in matrix X using formula (3), obtains matrix R；It is current according to mobile terminal Main business classification, selects the weight vector w={ w of the qos parameter of corresponding service classification in four kinds of types of service_d,w_dj, w_plr,w_gb,w_mb, calculate the judgement matrix V=wR after weighting and be shown below：

(3), after obtaining matrix V, positive ideal option A is calculated according to formula (5), (6) respectively⁺With ill ideal solution A^-；

Wherein,Different candidate network QoS each elements in for the judgement matrix V after weighting Maximum,It is the different candidate network QoS each elements in the judgement matrix after weighting for V Minima；

(4) distance of each candidate network and positive ideal scheme is calculated further according to formula (7) and formula (8)And with negative reason Think the distance of scheme

Wherein, v_ijFor the judgement matrix after the weighting of the corresponding j-th candidates network parameter values of i-th candidate network； For after weighting judgement matrix in j-th candidates network parameter element maximum,For jth in the judgement matrix after weighting The minima of individual candidate network param elements；

(5) finally candidate network N is drawn using formula (8)_iRelative to positive preferable option A⁺Degree of closeness

Compare each candidate networkValue, selectsThe maximum candidate network of value is used as objective network.Other steps And parameter is identical with one of specific embodiment one to three.

Specific embodiment five：Unlike one of present embodiment and specific embodiment one to four：In formula (1) each Element is the raw value of certain qos parameter in network of relation, i-th network N_iAll QoS parameter values row of matrix X' Vector { d_i',dj_i',plr_i',gb_i',mb_i' represent；Due to qos parameter be divided into the type of being the bigger the better (such as GB and MB) and it is more little more The value of the smaller the better type GB and MB parameters, for unification and variety trend, is asked reciprocal, is sentenced by good type (such as D, DJ and PLR) Certainly matrix X-type (2), wherein d_i=1d_i', dj_i=1/dj_i', plr_i=1/plr_i', gb=gb_i', mb=mb_i'.Other steps and Parameter is identical with one of specific embodiment one to four.

Embodiment one：

A kind of heterogeneous network system of selection distinguished based on type of service weights of the present embodiment, specifically according to following steps Prepare：

Step one, using the whole network configuration in Network finding and static information collection module (1) MIHF it is mobile eventually End decision-making is in current radio access network (RadioAccessNetwork, RAN), during communication, discovery adjacent R AN simultaneously collects phase The static information of adjacent RAN, and the static information for collecting adjacent R AN is stored in the local database；Wherein MIHF is complete to assume The configured Media Independent Handover function of IEEE802.21 of net (MediaIndependent HandoverFunction, MIHF)；

Step 2, the wireless static that mobile terminal is detected current RAN using candidate network qos parameter numerical value acquisition module State information signal intensity (RadioSignalStrength, RSS) drops to when needing switching threshold, then mobile terminal is to target Network initiates handover request, is switched to the wireless static information of current RAN in local data base, detects the wireless static of RAN successively The RSS of state information, will detect that adjacent R AN on communication threshold includes candidate network ranks, is collected by MIHF and is worked as The QoS parameter value of front candidate network ranks；

Step 3, by the main business classification of current mobile terminal decision-making, joined according to the QoS of current candidate network ranks Numerical value calculates qos parameter weight vector using AHP algorithms and evaluation algorithms calculate qos parameter weight vector and obtain candidate network Ideal approaches value, and the judgement matrix after being configured to weight using the QoS parameter value of current candidate network ranks selects ideal to force The maximum candidate network of close values, will select ideal to approach the maximum candidate network of value as objective network；Wherein, AHP algorithms meter The QoS weight vectors of four kinds of types of service are calculated, the QoS weight vectors of calculate four kinds of types of service are present in data base； Four kinds of types of service include flowing class, interactive class, background class and conversation class；

(1) assume that the candidate network number that mobile terminal determines is m, using the candidate network QoS parameter value construction for obtaining Judgement matrix X'；

Wherein, in formula (1), each element is the raw value of qos parameter in candidate network, i-th candidate network N_iInstitute There is QoS parameter value represent d with the row vector of matrix X'_i'dj_i'plr_i'gb_i'mb_i' judgement matrix X' normalization is returned Judgement matrix X after one change

D：Candidate network time delay, DJ：Candidate network delay variation, PLR：Candidate network packet loss, GB：Candidate network is ensured Bit rate, MB：Candidate network Maximum Bit Rate, i be i-th network, 1≤i≤m,

r_ijCorresponding QoS parameter value corresponding to i-th network, wherein, m is candidate network number, and j is that candidate network is joined Number；x_ijFor the element of the i-th row of X matrix jth row, Chinese implication is i-th network, j-th QoS parameter value normalized value；

2) it is in network of relation certain to adjudicate the element in matrix X to be standardized each element in formula (1) using formula (3) The raw value of individual qos parameter, i-th network N_iAll QoS parameter values can use matrix X' row vector { d_i',dj_i', plr_i',gb_i',mb_i' represent；Due to qos parameter can be divided into the type of being the bigger the better (such as GB and MB) and the smaller the better type (such as D, DJ and PLR), for unification and variety trend, the value of the smaller the better shape parameter is asked into reciprocal, obtain adjudicating matrix X-type (2), its Middle d_i=1/d_i', dj_i=1/dj_i', plr_i=1/plr_i', gb=gb_i', mb=mb_i'；Obtain matrix R；Worked as according to mobile terminal Front main business classification, selects the weight vector w={ w of the qos parameter of corresponding service classification in four kinds of types of service_d, w_dj,w_plr,w_gb,w_mb, calculate the judgement matrix V=wR after weighting and be shown below：

(3), after obtaining matrix V, positive ideal option A is calculated according to formula (5), (6) respectively⁺With ill ideal solution A^-；

Wherein,Different candidate network QoS each elements in for the judgement matrix V after weighting Maximum,It is the different candidate network QoS each elements in the judgement matrix after weighting for V Minima；

(4) distance of each candidate network and positive ideal scheme is calculated further according to formula (7) and formula (8)And with negative reason Think the distance of scheme

Wherein, v_ijFor the judgement matrix after the weighting of corresponding j-th network parameter values of i-th candidate network；To add The maximum of j-th candidates network parameter element in judgement matrix after power,For j-th time in the judgement matrix after weighting The minima of network selection network param elements；

(5) finally candidate network N is drawn using formula (8)_iRelative to positive preferable option A⁺Degree of closeness

Compare each candidate networkValue, selectsThe maximum candidate network of value is used as objective network；

Different service types are carried out under average weight algorithm and step analysis algorithm with network using Matlab and selects imitative Very；Three candidate networks N1 (UMTS), N2 (WiMAX), N3 (WLAN), it is assumed that mobile terminal has detected that current network are set RSS is less than QoS of survice demand threshold, will switch, and the current main business type of mobile terminal is the four of 3GPP definition One of which in type, detects three candidate networks Nl, N2 and N3, obtains three times by the MIHF of IEEE802.21 The current QoS parameter value of network selection network, is utilized respectively the average weight algorithm TOAWA of service-oriented classification (per class QoS of survice parameter Weights it is equal) and the present invention in the QoS weights methods of utilization chromatographic analysises algorithm differentiated service type that propose, wait currently The objective network selected under the conditions of selecting network QoS, records selection result；Switching frequency is l0 ten thousand times, wherein every kind of business Type is each 2.5 ten thousand times, the one after the other successively of every kind of type of service；

As shown in figure 4, the average all kinds of QoS of survice weights of traditional method, will cause a large amount of access networks N1 of business and N2, and Cause network N 3 idle；As shown in figure 5, differentiated service QoS weights in the present invention, are the most suitable network of each class service selection Access, so as to equilibrium assignment is per class business, effectively utilizes Internet resources；

The objective network percentage ratio situation of every class service selection of traditional method and this method is respectively as shown in Figure 4 and Figure 5； Average weight algorithm more meets people N1 (UMTS) by conversation class and stream class service, is so likely to cause N1 long-term blockages, drop Low terminal connects the handover success rate of people N1；In addition, allowing exchange class and background class traffic more to meet people N3 (WLAN), it is likely that increase The collision probability of big business datum, reduces data transmission success；And set forth herein based on analytic hierarchy process (AHP) network select Algorithm is by Streaming Media class and exchanges the net that class business is more switched to this relative High Data Rates of N2 (WiMAX), low error rate Network, while QoS of survice is ensured, reasonably selection target network such as Fig. 6；Meanwhile, compared with average weight algorithm, Hou Zhe On the premise of ensureing QoS of survice demand, the load of three candidate networks is effectively balanced.

Claims (2)

1. a kind of heterogeneous network system of selection distinguished based on type of service weights, it is characterised in that：It is a kind of to be based on type of service The heterogeneous network system of selection that weights are distinguished comprises the steps：

Step one, the MIHF configured using the whole network in Network finding and static information collection module (1) are determined in mobile terminal When plan is to communicate in current radio access network RAN, finds adjacent R AN and collect the static information of adjacent R AN, and phase will be collected The static information of adjacent RAN is stored in the local database；

Step 2, the wireless static letter that mobile terminal is detected current RAN using candidate network qos parameter numerical value acquisition module Information signal intensity RSS drops to when needing switching threshold, then mobile terminal initiates handover request to objective network, is switched to local The wireless static information of current RAN in data base, detects the RSS of the wireless static information of RAN successively, will detect in communication Adjacent R AN on threshold value includes candidate network ranks, collects the QoS parameter value of candidate network ranks by MIHF；

Step 3, by the main business classification of current mobile terminal decision-making, utilized according to the QoS parameter value of candidate network ranks AHP algorithms calculating qos parameter weight vector and evaluation algorithms calculating qos parameter weight vector obtain the ideal of candidate network and approach Value, it is maximum that the judgement matrix after being configured to weight using the QoS parameter value of current candidate network ranks selects ideal to approach value Candidate network, by select ideal approach value maximum candidate network as objective network；Wherein, four kinds of types of service include stream Media class, interactive class, background class and conversation class；A kind of heterogeneous network selecting party distinguished based on type of service weights is completed Method；

In step 3 by the selection preferable maximum candidate network of value of approaching as objective network detailed process it is：

(1) assume that the candidate network number that mobile terminal determines is m, using the candidate network QoS parameter value construction judgement for obtaining Matrix X'；

Wherein, i-th candidate network N_iThe row vector of all QoS parameter value matrix X' represent d_i'dj_i'plr_i'gb_i'mb_i' Judgement matrix X' normalization is obtained into the judgement matrix X after normalization；

D：Candidate network time delay, DJ：Candidate network delay variation, PLR：Candidate network packet loss, GB：Candidate network ensures bit Rate, MB：Candidate network Maximum Bit Rate, i：I-th network, 1≤i≤m；

$\begin{array}{cc}{r}_{ij}=x_{ij}/\sqrt{\underset{i=1}{\overset{m}{\Σ}}{x}_{ij}^2};& j=D,DJ,PLR,GB,MB\end{array}---\left(3\right)$

r_ijCorresponding QoS parameter value corresponding to i-th network, wherein, m is candidate network number, and j is X for candidate network parameter Matrix column number；x_ijFor the element of the i-th row of X matrix jth row, j-th QoS parameter value normalized value of as i-th network；

(2) it is standardized to adjudicating the element in matrix X using formula (3), obtains matrix R；According to the current master of mobile terminal Class of service is wanted, the weight vector w={ w of the qos parameter of corresponding service classification are selected in four kinds of types of service_d,w_dj,w_plr, w_gb,w_mb, calculate the judgement matrix V=wR after weighting and be shown below：

$\begin{array}{c}V=\left[\begin{array}{ccccc}{v}_{1d}& v_{1dj}& v_{1plr}& v_{1gb}& v_{1mb}\\ .& .& .& .& .\\ .& .& .& .& .\\ .& .& .& .& .\\ v_{id}& v_{idj}& v_{iplr}& v_{igb}& v_{imb}\\ .& .& .& .& .\\ .& .& .& .& .\\ .& .& .& .& .\\ v_{md}& v_{mdj}& v_{mplr}& v_{mgb}& v_{mmb}\end{array}\right]\\ =\left[\begin{array}{ccccc}{w}_{1d}{r}_{1d}& w_{1dj}{r}_{1dj}& w_{1plr}{r}_{1plr}& w_{1gb}{r}_{1gb}& w_{1mb}{r}_{1mb}\\ .& .& .& .& .\\ .& .& .& .& .\\ .& .& .& .& .\\ w_{1d}{r}_{1d}& w_{idj}{r}_{idj}& w_{iplr}{r}_{iplr}& w_{igb}{r}_{igb}& w_{imb}{r}_{imb}\\ .& .& .& .& .\\ .& .& .& .& .\\ .& .& .& .& .\\ w_{1d}{r}_{1d}& w_{mdj}{r}_{mdj}& w_{mplr}{r}_{mplr}& w_{mgb}{r}_{mgb}& w_{mmb}{r}_{mmb}\end{array}\right]\end{array}---\left(4\right)$

(3), after obtaining matrix V, positive ideal option A is calculated according to formula (5), (6) respectively⁺With ill ideal solution A^-；

$\{v_d^{+},v_{dj}^{+},v_{plr}^{+},v_{bg}^{+},v_{mb}^{+}\}---\left(5\right)$

$A^{-}=\left\{\underset{1\≤i\≤m}{min}{v}_{ij}\right|(j=d,dj,plr,gb,mb)\}=\{v_d^{-},v_{dj}^{-},v_{plr}^{-},v_{gb}^{-},v_{mb}^{-}\}---\left(6\right)$

Wherein,The maximum of the different candidate network QoS each elements in for the judgement matrix V after weighting Value,It is the minimum for adjudicating the different candidate network QoS each elements in matrix after weighting for V Value；

(4) each candidate network is calculated with positive ideal scheme apart from S further according to formula (7) and formula (8)_i ⁺And and ill ideal solution Apart from S_i ^-；

$\begin{array}{ccc}{S_i}^{+}=\sqrt{\underset{j}{\Σ}{(v_j^{+}-v_{ij})}^2};& j=1,2,\mathrm{...},m& j=d,dj,plr,gb,mb\end{array}---\left(7\right)$

$\begin{array}{ccc}{S_i}^{-}=\sqrt{\underset{j}{\Σ}{(v_j^{-}-v_{ij})}^2};& j=1,2,...,m& j=d,dj,plr,gb,mb\end{array}---\left(8\right)$

Wherein, v_ijFor the judgement matrix after the weighting of the corresponding j-th candidates network parameter values of i-th candidate network；To add The maximum of j-th candidates network parameter element in judgement matrix after power,For j-th candidates in the judgement matrix after weighting The minima of network parameter element；

(5) finally candidate network N is drawn using formula (8)_iRelative to positive preferable option A⁺Degree of closeness C_i ^*；

$\begin{array}{ccc}{C}_i^{*}=S_i^{-}/(S_i^{+}+S_i^{-});& 0\≤C_i^{*}\≤1,& i=1,2,...,m\end{array}---\left(9\right)$

Select C_i ^*The maximum candidate network of value is used as objective network.

2. a kind of heterogeneous network system of selection distinguished based on type of service weights according to claim 1, its feature are existed In：To the process that the element in judgement matrix X is standardized it is：

In formula (1), each element is the raw value of certain qos parameter in network of relation, i-th network N_iAll qos parameters Row vector { d of the value with matrix X'_i',dj_i',plr_i',gb_i',mb_i' represent；The value of GB and MB parameters is asked into reciprocal, is obtained Judgement matrix X-type (2), wherein d_i=1/d_i', dj_i=1/dj_i', plr_i=1/plr_i', gb=gb_i', mb=mb_i'。