CN101835235A - Cognitive-Based Routing Method for Heterogeneous Networks - Google Patents

Abstract

The invention discloses a routing method for a heterogeneous network based on cognition, mainly solving the problem that the existing method considers only part of environment factors of the network, and neglects the dynamic complexity of a heterogeneous network environment so as not to cognize and predict the environment. The method comprises the following steps: firstly, judging the types of the existing available networks, perceiving a network environment state, evaluating the network environment state by using a Q-learning method and calculating end-to-end time delay and the estimated values thereof; secondly, using a utility function based on multiple parameters to obtain the utility values of various networks according to the perceived network environment state; then selecting the network with the biggest utility value by a probability of P=0.9, and selecting other available networks as target networks by the probability of 1-P; and finally continuously transmitting businesses, implementing network switch and updating the network environment state by the target networks. The routing method improves the utilization ratio of network resources, achieves load balance of the network and can be used in the heterogeneous network environment.

Description

Cognitive-Based Routing Method for Heterogeneous Networks

technical field

The invention belongs to the technical field of wireless communication, and relates to a routing method, in particular to a cognitive-based heterogeneous network routing method in a heterogeneous network, which is used for a mobile terminal in a wireless network to select a suitable network in an area covered by multiple networks. path transmission business.

Background technique

With the continuous development of communication technology, since the current spectrum resources are fixedly allocated, there are fewer and fewer spectrum resources available in the entire spectrum space, and the uneven allocation of spectrum resources is one of the problems that need to be solved urgently. Since the current network environment is restricted by many factors such as the shortage of spectrum resources and the constraints of radio wave transmission characteristics, it cannot continue to expand. Therefore, it is necessary to study a new network that can make full use of various network resources and ensure the improvement of network service quality. technology.

The trend of the future network is to gradually move towards cognition, integration, and ubiquity, that is, cognitive network. A cognitive network is a network system with a cognitive process that perceives current network conditions and makes plans and decisions to make further actions according to the conditions. It can learn and predict the network environment and make subsequent decisions. At the same time, the convergence of various networks is also the trend of future network development. The convergence of heterogeneous networks has many advantages, such as improving network scalability, making full use of network resources, and better meeting user needs. Therefore, it is necessary to design a reasonable heterogeneous network routing method to solve the above problems, that is, it is hoped to select a network user that is most suitable for the current user to transmit services in overlapping coverage areas of various networks. In a heterogeneous network environment with cognitive capabilities, how to make rational use of network resources and meet the needs of both users and the network as much as possible, and how users can choose a suitable network to transmit services through their own cognitive capabilities, that is, to perform network switching, so as to To achieve the improvement of network performance is the difficulty of routing algorithm.

At present, there are many studies on heterogeneous network selection methods. The existing algorithms are mainly based on fuzzy logic algorithms, algorithms based on certain factors such as RSS, power, delay and other QoS requirements, network selection algorithms based on utility, and algorithms based on Decision-making methods for determining parameter weighting, etc. These algorithms only consider some environmental factors of the network, and do not consider the interaction between user preferences, user types and network environment, and also ignore the dynamic and complex environmental characteristics of heterogeneous networks, which cannot be well adapted to heterogeneous networks. Complex heterogeneity and dynamic change, and the network environment cannot be recognized and predicted, and the requirements of future network development trends cannot be met.

Contents of the invention

The purpose of the present invention is to overcome the deficiencies of the above-mentioned prior art, and provide a cognitive-based heterogeneous network routing method to adapt to the complexity and dynamic variability of heterogeneous networks, comprehensively considering the relationship between user preferences, user types and network environments The interaction of the environment, cognition and prediction of the environment, to meet the needs of future network development.

To achieve the above object, the routing method of the present invention comprises the following steps:

(1) The mobile terminal periodically obtains the currently available network types, perceives the network environment status, uses the Q-learning method to estimate the network environment status, obtains the estimated value of each network environment status, and calculates the end-to-end delay and its estimated value;

(2) According to the obtained network environment status, the mobile terminal performs normalization processing on the available network bandwidth f _cap , the network usage fee f _pric and the end-to-end delay T _total to obtain the end-to-end delay, the available network bandwidth and the network Using the weighted value of the cost, the network with the highest utility value is selected by the following function:

Max U _i =F _i (f _cap , f _pric , T _total )

Among them, F _i (f _cap , f _pric , T _total )=ω ₁ (BW _max -f _cap )+ω ₂ f _pric +ω ₃ T _total is a utility function based on multiple parameters, and ω ₁ is the remaining available bandwidth of the network Weight value, ω ₂ is the weight value of the network usage fee, ω ₃ is the weight value of the end-to-end delay, and the above weight values are obtained by the analytic hierarchy process;

(3) According to the obtained network with the largest utility value, the mobile terminal selects the network with probability P=0.9, and selects other available networks with probability 1-P as the target network for service transmission;

(4) The mobile terminal performs service transmission according to the obtained target network, and judges whether the target network is the same as the current network. If they are the same, the mobile terminal directly uses the network to transmit services; otherwise, the mobile terminal sends a switching request to switch from the current network To the target network, continue to transmit the service through the target network, and update the network environment status and the estimated value of each state at the same time. Ensure the optimization of service transmission path.

Compared with the prior art, the present invention has the following advantages:

1) The present invention uses the cognitive ability of the environment during the service transmission process of the mobile terminal to predict and estimate the state of the network environment, so that the state of the environment obtained by the mobile terminal is closer to the real situation, instead of blindly believing in the network environment State historical information, and can predict the environmental state trend of the network for a period of time in the future, which is beneficial for mobile terminals to make choices in line with network resource allocation, reduces network congestion, and is also obtained in network resource allocation and network resource utilization. Very good improvement and enhancement;

2) The present invention reduces the excessive dependence of the mobile terminal on a single environmental factor because the utility function combined with multiple parameters is used as the criterion for selecting a network in the mobile terminal, and is more in line with the overall network environment where the mobile terminal is located, so that the mobile terminal Choose a network that is more in line with your own requirements for business transmission, which can achieve the balance of network resources and improve the utilization of network resources;

3) Since the present invention uses a certain probability to select a network for transmitting services at the mobile terminal, by reducing the probability that the mobile terminal always selects a certain network with the best utility, the network congestion caused by this is reduced, thereby reducing the traffic congestion. The packet loss rate improves the probability of successful service transmission; it also increases the utilization rate of some suboptimal networks, thereby improving the resource utilization rate.

The simulation results show that, compared with the existing methods, the present invention improves the utilization rate of network resources and the success probability of service transmission, reduces the end-to-end time delay and packet loss rate of service transmission, realizes network load balancing, and improves network efficiency. communication performance.

Description of drawings

Fig. 1 is the basic application scenario figure that the present invention uses;

Fig. 2 is a flow chart of the heterogeneous network routing method of the present invention;

Fig. 3 is a scene topology diagram used in the present invention;

Fig. 4 is the simulation diagram of the packet loss rate obtained by the existing network routing method;

Fig. 5 is the simulation figure of the network successful transmission ratio that existing network routing method obtains;

FIG. 6 is an end-to-end time delay simulation diagram obtained by an existing network routing method;

Fig. 7 is the simulation diagram of the packet loss rate obtained by the network routing method proposed by the present invention;

Fig. 8 is the simulation diagram of the network successful transmission ratio that the network routing method that the present invention proposes obtains;

FIG. 9 is a simulation diagram of end-to-end time delay obtained by the network routing method proposed by the present invention.

Detailed ways

In order to clearly illustrate the method in the present invention, an application scenario diagram and a flow chart are given below and detailed descriptions are given in conjunction with the accompanying drawings.

The present invention is applied in the application scenario of heterogeneous network as shown in Fig. 1, and this scenario is in a network partially overlapped and covered by wireless local area network WLAN, general packet radio network GPRS and wideband wireless access network WiMAX area, the current mobile terminal can choose any network that has covered itself to transmit services during the service transmission process, rather than being limited to a specific network; the mobile terminal needs to find a network that is more in line with the current network environment and satisfies During the whole process, the mobile terminal needs to select one of the networks for service transmission among various networks, and constantly predict the state of the network environment, so as to obtain an optimal path for service transmission.

Referring to Fig. 2, the routing steps of the present invention are as follows:

Step 1, the mobile terminal periodically acquires the currently available network set Net(i)={N(j), j=1~n} and its corresponding network environment status, where the network types include wireless local area network WLAN, general packet wireless network GPRS and wideband wireless access network WiMAX, the network currently used by mobile terminals is N(i), the network environment status includes network available bandwidth f _cap , network usage fee f _pric and data transmission rate rat(N(i), N( j)), switching delay sw(N(i), N(j)) and access delay ac(N(i), N(j)); the initial value of the network serial number net_num with the largest utility value is set to 0, delay estimate The initial setting is 30, and the maximum utility value max is initially set to 0.

Step 2. According to the obtained network environment state, the mobile terminal adopts the Q-learning method with cognitive ability to predict and estimate the network environment state. According to the following formula, obtain the estimated value of each state, and calculate the required end-to-end delay T _total and estimated delay

rat_e(N(i), N(j)) = rat(N(i), N(j)) + λ _rat [rat_e(N(i), N(j)) - rat(N(i), N (j))]

ac_e(N(i), N(j))=ac(N(i), N(j))+λ _ac [ac_e(N(i), N(j))-ac(N(i), N (j))]

sw_e(N(i), N(j))=sw(N(i), N(j))+λ _sw [sw_e(N(i), N(j))-sw(N(i), N (j))]

${\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; total</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; sw</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; ac</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; pk</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; rat</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; Net</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; min</span>}}\mathrm{<span\; class="notranslate">\; Qt</span>}\stackrel{<span\; class="notranslate">\; ^</span>}{<span\; class="notranslate">\; \_</span>}{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; Q</span>}\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; \_</span>}{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; \&lambda;Q</span>}\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; \_</span>}{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; Net</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; max</span>}}\mathrm{<span\; class="notranslate">\; Q</span>}\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; \_</span>}{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>$

为业务从网络N(j)通过网络N(k)传输的端到端时延估计值，λ为Q值的估计速度，r_ij为业务从网络N(j)传输到网络N(j)所获得的效用值。Among them, rat_e(N(i), N(j)) is the estimated value of the data transmission rate, λ _rat is the estimated speed of the data transmission rate, sw_e(N(i), N(j)) is the slave network N(i) The estimated value of handover delay required to switch to network N(j), λ _sw is the estimated speed of handover delay, ac_e(N(i), N(j)) is the estimated value of access delay, λ _ac is the access delay The estimated speed of the input delay, pk is the size of the business to be transmitted,

is the estimated value of the end-to-end delay of the service from the network N(j) to the network N(k), λ is the estimated speed of the Q value, r _ij is the transmission time of the service from the network N(j) to the network N(j) The utility value obtained.

Step 3: According to the obtained network environment status, the mobile terminal normalizes the available network bandwidth, network usage fee and end-to-end delay by the following formula:

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; cap</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; cap</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; BW</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}{{\mathrm{<span\; class="notranslate">\; BW</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; BW</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}$

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; pric</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; pric</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; price</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}{{\mathrm{<span\; class="notranslate">\; price</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; price</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}$

${\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; total</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; total</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}{{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}}$

Among them, BW _min is the minimum available bandwidth available to the network, BW _max is the maximum available bandwidth available to the network, Price _min is the minimum fee to be paid for network use, Price _max is the maximum fee to be paid for network use, and T _min is the service transmission T _max is the maximum end-to-end delay allowed by service transmission.

Step 4, according to the obtained normalized value of end-to-end delay, available network bandwidth and network usage cost, the mobile terminal adopts AHP to obtain the weight value of end-to-end delay, network available bandwidth and network usage cost, according to The different relative influence degrees of the three are assigned to the degree of influence. The value of each level can be randomly selected from the integers within 1 to 9. The higher the degree of influence, the smaller the selected value; otherwise, the lower the degree of influence, the choice The larger the value of , assuming that the first randomly selected influence degree levels are 1, 2, 5, a comparison matrix W composed of influence degree levels is obtained:

$\mathrm{<span\; class="notranslate">\; W</span>}<span\; class="notranslate">\; =</span>\left[\begin{array}{ccc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 5</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 5</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]<span\; class="notranslate">\; .</span>$

Step 5, according to the obtained comparison matrix W, calculate the maximum eigenvalue λ _max by |W-λ _max E|=0, where E is a 3×3 identity matrix; then use the consistency index formula CI=(λ _max - n)/(n-1) to obtain CI, where n is the dimension of the comparison matrix W; and according to the obtained consistency index CI, CR is obtained by the consistency ratio index formula CR=CI/RI, where RI is random Consistency index, different n values have fixed RI values, which can be obtained by looking up the random consistency index table.

Step 6 _, judge whether the matrix is acceptable according to the obtained CR. When CR<0.1, it is considered that the comparison _matrix W is acceptable, and the eigenvector x=(x ₁ , x ₂ , x ₃ ) ^T , take the obtained x ₁ , x ₂ , x ₃ values as the weight values ω ₁ , ω ₂ , ω ₃ of delay, bandwidth and cost, and complete the process of solving the weight values of each state; otherwise , if the comparison matrix W is considered unacceptable, proceed to step 3 to reselect an acceptable comparison matrix, and then calculate the weight value.

Step 7, according to the obtained weight value, use a multi-parameter-based utility function to synthesize the utility of various parameters for routing selection, select the network with the largest utility value and store the corresponding network serial number in net_num through the following functional formula:

Max U _i =F _i (f _cap , f _pric , T _total )

F _i (f _cap ，f _pric ，T _total )＝ω ₁ (BW _max -f _cap )+ω ₂ f _pric +ω ₃ T _total

Among them, ω ₁ is the weight value of the remaining available bandwidth of the network, BW _max is the total bandwidth allocated by the network, ω ₂ is the weight value of the network usage cost, and ω ₃ is the weight value of the end-to-end delay.

Step 8: According to the obtained network with the largest utility value, the mobile terminal randomly selects a real number in the interval [0, 1] as the probability P of selecting a network. If P<0.9 holds true, the mobile terminal selects the network with the largest utility value as the transmission The target network of the service, otherwise, randomly select any available network in the set Net(i)={N(j), j=1~n} as the target network.

Step 9: The mobile terminal performs service transmission according to the obtained target network, and judges whether the target network is the same as the current network. If they are the same, the mobile terminal directly uses the network to transmit services; otherwise, the mobile terminal sends a switching request to switch from the current network to the target network, and continue to transmit the service through the target network.

Step 10, after the mobile terminal completes the above service transmission, re-update the network and its environmental state parameters, that is, use the switched network as the network currently used by the mobile terminal, and obtain new environmental state parameters under the network, and continue to select In the network for service transmission, when the service is transmitted to the backbone network, the mobile terminal also performs the same behavior to the nodes in the backbone network to ensure the optimization of the service transmission path.

Effect of the present invention can be further illustrated by following simulation:

Simulation conditions: use MATLAB simulation environment.

Fig. 3 is a topological diagram of the application scenario of the method of the present invention, which shows that the mobile terminal needs to send services from the source network to the destination network, wherein node 1 is the source network, node 9 is the destination network, and the mobile terminal can freely Access to any network transmission service. The network environment includes WLAN, WiMAX, GPRS networks and their corresponding backbone networks. Different types of networks have different data transmission rates, access delays, allocated network available bandwidth, and network usage fees. The network currently used by the mobile terminal is a WLAN network.

As shown in Figure 4, Figure 5 and Figure 6, in the existing network routing method, due to the constraint that resources cannot be shared between different types of networks, the arrival of a large number of services will cause the current network to be in a state of congestion and saturation for a long time, so Arriving business cannot be served immediately, causing the timer to time out, causing the packet loss rate to increase rapidly, the successful transmission rate of business in the network is greatly reduced due to a large number of packet loss, and the business waiting for service is also forced to stop due to network congestion Waiting for transmission, so the end-to-end delay also increases rapidly; at the same time, because the resources between the networks cannot be shared, the uneven service arrival between the networks makes the load unbalanced among the networks, and the resource utilization rate is relatively low.

As shown in Figure 7, Figure 8 and Figure 9, using the method proposed by the present invention, the mobile terminal has the ability to recognize the network environment, and can make corresponding effective estimates and predictions for the network by perceiving the state of the network environment, and then According to the predicted situation, a multi-parameter utility function is used as a criterion for selecting a network, and a network for service transmission is selected with a certain probability, and the network is not necessarily the currently used network. Therefore, when a large amount of business arrives, due to the cognitive ability of the mobile terminal, the selected network can know the congestion of the network, so the business will not be blocked in a large amount, and the packet loss rate will only increase slightly. At the same time, the successful transmission rate of the network It will not be greatly reduced due to the arrival of a large number of services, and the end-to-end delay will only increase slowly and then decrease to a certain value before there will be no major fluctuations. It can be seen that the method of the present invention can better realize load balancing of network resources, improve utilization rate of network resources, and improve network performance.

Claims (4)

1. A cognitive-based heterogeneous network routing method, comprising the steps of: (1) The mobile terminal periodically obtains the currently available network types, perceives the network environment status, uses the Q-learning method to estimate the network environment status, obtains the estimated value of each network environment status, and calculates the end-to-end delay and its estimated value; (2) According to the obtained network environment status, the mobile terminal performs normalization processing on the available network bandwidth f _cap , the network usage fee f _pric and the end-to-end delay T _total to obtain the end-to-end delay, the available network bandwidth and the network Using the weighted value of the cost, the network with the highest utility value is selected by the following function: Max U _i =F _i (f _cap , f _pric , T _total ) Among them, F _i (f _cap , f _pric , T _total )=ω ₁ (BW _max -f _cap )+ω ₂ f _pric +ω ₃ T _total is a utility function based on multiple parameters, and ω ₁ is the remaining available bandwidth of the network Weight value, ω ₂ is the weight value of the network usage fee, ω ₃ is the weight value of the end-to-end delay, and the above weight values are obtained by the analytic hierarchy process; (3) According to the obtained network with the largest utility value, the mobile terminal selects the network with probability P=0.9, and selects other available networks with probability 1-P as the target network for service transmission; (4) The mobile terminal performs service transmission according to the obtained target network, and judges whether the target network is the same as the current network. If they are the same, the mobile terminal directly uses the network to transmit services; otherwise, the mobile terminal sends a switching request to switch from the current network To the target network, continue to transmit the service through the target network, and update the network environment status and the estimated value of each state at the same time. Ensure the optimization of service transmission path. 2. The network selection and routing method according to claim 1, wherein the network environment status described in step (1) includes available network bandwidth _fcap , network usage fee _fpric and data transfer rate rat(N(i), N (j)), handover delay sw(N(i), N(j)) and access delay ac(N(i), N(j)). 3. The network selection routing method according to claim 1, wherein the perceived network environment state described in step (1) is carried out in the following steps: (3a) The mobile terminal acquires all currently available network types, represented by the set Net(i), Net(i)={N(j), j=1~n}, the network types include wireless local area network WLAN, general packet wireless network GPRS and broadband wireless access network WiMAX; (3b) For the network type, obtain its corresponding network environment status, and obtain the estimated value of the data transmission rate, access delay and handover delay through the following formula: rat_e(N(i), N(j)) = rat(N(i), N(j)) + λ _rat [rat_e(N(i), N(j)) - rat(N(i), N (j))] ac_e(N(i), N(j))=ac(N(i), N(j))+λ _ac [ac_e(N(i), N(j))-ac(N(i), N (j))] sw_e(N(i), N(j))=sw(N(i), N(j))+λ _sw [sw_e(N(i), N(j))-sw(N(i), N (j))] Among them, λ _rat is the estimated speed of data transmission rate, rat_e(N(i), N(j)) is the estimated value of data transmission rate, λ _ac is the estimated speed of access delay, ac_e(N(i), N(j)) is the estimated value of the access delay, λ _sw is the estimated speed of the handover delay, sw_e(N(i), N(j)) is the estimated value of the handover delay; (3c) According to the estimated value obtained above, the end-to-end time delay T _total required by the mobile terminal during service transmission is obtained by the following formula: ${\mathrm{<span\; class="notranslate">\; T</span>}}_{\mathrm{<span\; class="notranslate">\; total</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; sw</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; ac</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; pk</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; rat</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; Net</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; min</span>}}\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Qt</span>}<span\; class="notranslate">\; \_</span>{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>$ Among them, pk is the size of the service, pk/rat_e(N(i), N(j)) is the time required for the service to be transmitted between the networks,

Estimated value of the end-to-end delay required for the transmission of traffic from network N(j) through network N(k); (3d) The mobile terminal uses the Q-learning method to estimate the obtained environmental state, and obtains the estimated value of the end-to-end delay of the service from the network N(i) to the network N(j): $\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Qt</span>}<span\; class="notranslate">\; \_</span>{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; \&lambda;</span>}\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Qt</span>}<span\; class="notranslate">\; \_</span>{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&lambda;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; [</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; Net</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; max</span>}}\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Qt</span>}<span\; class="notranslate">\; \_</span>{\mathrm{<span\; class="notranslate">\; e</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; ]</span>$ Among them, λ is the estimated speed of Q value, and r _ij is the utility value obtained from the service transmission from network N(i) to network N(j). 4. The network selection routing method according to claim 1, wherein the state weight value ω ₁ described in step (2), ω ₂ , ω ₃ is obtained by the following steps: (4a) According to the influence degree of delay, bandwidth and cost relative to the mobile terminal, assign three different influence degree grades, the value of each grade can be randomly selected from integers within 1 to 9, the higher the influence degree grade, The smaller the selected value, on the contrary, the lower the level of influence, the larger the selected value. Assuming that the first random selection of the level of influence is 1, 2, 5, a comparison matrix W composed of the level of influence is obtained: $\mathrm{<span\; class="notranslate">\; W</span>}<span\; class="notranslate">\; =</span>\left[\begin{array}{ccc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 5</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 2</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 5</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]<span\; class="notranslate">\; ;</span>$ (4b) According to the obtained comparison matrix W, the maximum eigenvalue λ _max is obtained by |W-λ _max E|=0, wherein E is a 3×3 identity matrix; (4c) According to the maximum eigenvalue λ _max obtained, the CI is obtained by the consistency index formula CI=(λ _max -n)/(n-1), and the consistency check is performed with this index, where n is the comparison matrix W the dimension; (4d) According to the obtained consistency index CI, CR is obtained by the consistency ratio index formula CR=CI/RI, where RI is a random consistency index, and different n values have fixed RI values, which can be found by searching the random consistency The index table is obtained; (4e) When CR<0.1, it is considered that the comparison matrix W is acceptable, then the eigenvector x=(x ₁ , x ₂ , x ₃ ) ^T corresponding to the largest eigenvalue λ _max is obtained by Wx=λ _max , and The x ₁ , x ₂ , x ₃ values of each state are used as the weight values ω ₁ , ω ₂ , ω ₃ of each state, and the process of solving the weight values of each state is completed; otherwise, it is considered that the comparison matrix W is unacceptable, and then proceed to step (4a) Reselect an acceptable comparison matrix, and then calculate the weight value.

Images