CN103795610B - The instant messaging dispatching method of a kind of Consumer's Experience justice and device - Google Patents

Abstract

The invention discloses a method for dispatching instant message services with fair user experience, which includes the following steps: (1) calculating the attention increment of each sub-service in the current period according to the normalized quality score of each sub-service; according to the attention degree Increment and the user attention of the previous period, calculate the user attention of the current period; (2) Select the sub-service with the highest user attention in the current period as the representative service of the user, and let the normalized quality score of the representative service be used as the user’s representative service. The representative service quality score of ; (3) Scheduling the sub-service representing the user with the worst service quality score. An instant message service scheduling device includes a data queue delay information maintenance module, a user experience estimation module and a sub-service data flow scheduling module. The present invention can estimate user experience in the scenario where there are multiple instant message users in the network and each user uses one or more sub-services, and realize user experience fairness through data flow scheduling.

Description

A method and device for scheduling instant message services with fair user experience

technical field

The invention relates to the communication field, in particular to a method for dispatching instant message services in a wireless network and an implementation device thereof.

Background technique

The instant messaging service is one of the most widely used services in the current mobile Internet. The number of its users is huge, taking Skype as an example, its number of simultaneous online users is as high as 70 million. The instant message service refers to a mixed special service including three instant message sub-services (hereinafter referred to as sub-services) such as video conversation, voice conversation and text. Different from general data services (such as ftp and email), instant message services are conversational interactive services with real-time requirements. However, unlike real-time services that can obtain dedicated channels or bandwidth guarantees after being admitted, users of instant messaging services need to compete for shared channel bandwidth, and there is no quality assurance mechanism. Therefore, scheduling technology is particularly important for instant messaging service quality and user experience.

In a network scenario where there are multiple instant messaging users at the same time, it is a technical challenge to design scheduling to enable users to obtain similar service experience. The reasons are as follows:

(1) User experience is not only affected by the number of network resources and users in the network, but also by the type and number of sub-services used by users;

(2) Users can interact through multiple sub-services, but different types of sub-services have different impacts on the same user, and the same sub-service has different impacts on different users, making it difficult to measure and compare user experience.

(3) Sub-services have a stepped demand for system bandwidth. Users using different types of sub-services have different requirements for bandwidth. The characteristics of the same user paying attention to different sub-services at different times lead to dynamic changes in their requirements for system bandwidth. For such dynamic and differentiated requirements, a simple scheduling method based on resource fairness cannot achieve a fair user experience.

In the document "Delay Sensitive Scheduling Schemes for Heterogeneous QoS over Wireless Networks", Z. Han et al. present a scheduling method that can provide fair services for delay-sensitive service users. It proposes User Satisfaction Factor (USF) as a measure of user benefits (experience). However, this USF is designed for the scenario where each user only uses one sub-service at any time, and is not suitable for the instant message scenario where each user may use multiple sub-services. In the document "Low Complexity Cross-Layer Design with PacketDependent Scheduling for Heterogeneous Traffic in Multiuser OFDM Systems", N.Zhou et al. proposed an optimization goal with maximum weighted sum capacity (MWSC) of the user, and the data packet The dispatch method for the dispatch object. However, the weight in the objective function of this method is only related to the sub-service delay, and cannot reflect the impact of factors such as the type of sub-service and user attention on user experience. When an instant message user uses multiple sub-services, this method will limit Resources allocated to certain sub-services that have little impact on user experience, resulting in the fact that the sub-services that users really care about cannot be guaranteed.

The US patent No. US8,315,213 discloses a technology for providing proportional fair services for users using multiple mobile services, claiming that it can be applied to instant messaging services. It allocates sub-channel resources according to sub-channel quality and average throughput of mobile terminals. However, the throughput cannot reflect the quality of the three real-time sub-services of the instant message service. This technology is not suitable for the instant message service scenario, because the main factor affecting the user experience of the instant message service is the sub-service delay. US Patent No. US7,957,417 discloses a technology that can provide fair services to mobile service users (including instant message users) in a multi-access wireless network. The method is to set a high priority for users who use less resources based on the average resource usage of current users and the forecast of future resource usage, so that the number of users who receive the service is the largest, but the experience of some users who use more resources will always be poor. .

Contents of the invention

The purpose of the present invention is to overcome the deficiencies of the prior art, aim at the characteristics of the instant message service itself, and provide an instant message service dispatching method and device with fair user experience according to the analysis of user behavior characteristics.

A method for dispatching instant message services with fair user experience, comprising the steps of:

(1) Calculate the normalized quality score of each sub-service run by the user in the current period according to the average delay of the sub-service data queue;

(2) According to the normalized quality score of each sub-service in the current period, calculate the attention increment of each sub-service in the current period; according to the user attention of each sub-service in the previous period and the attention increment of each sub-service in the current period , to calculate the user attention of each sub-service in the current period;

(3) Select the sub-service with the highest user attention in the current period as the user's representative service, and make the normalized quality score of the representative service as the user's representative service quality score;

(4) Traverse all users within the coverage of the base station according to steps (1) to (3), get the representative service quality score of each user, select the user with the worst representative service quality score as the user to be served, and take the user attention as Probabilistically select a sub-service of the user to be served for service, so as to improve the normalized quality score of the sub-service of the user to be served.

When the step (4) is serving the sub-service of the selected user to be served, if the sub-service includes two data streams, the data stream with a larger average time delay is served.

The formula for calculating the normalized quality score of each sub-business in step (1) is as follows:

${\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; max</span>}<span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; AD</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; MS</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \}</span><span\; class="notranslate">\; ,</span>& <span\; class="notranslate">\; *</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; V</span>}\\ \mathrm{<span\; class="notranslate">\; 93.2</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.02</span>}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 0.11</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 177.3</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 177.3</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& <span\; class="notranslate">\; *</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; A</span>}\\ \mathrm{<span\; class="notranslate">\; 4.73</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; ln</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; 10</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3</span>}}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; T</span>}}<span\; class="notranslate">\; )</span>& <span\; class="notranslate">\; *</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; T</span>}\end{array}\right.$

$\mathrm{<span\; class="notranslate">\; AD</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 3.92</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3.24</span>}<span\; class="notranslate">\; \&times;</span>{\mathrm{<span\; class="notranslate">\; 10</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; +</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; I</span>}}<span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; MS</span>}<span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1.38</span>}<span\; class="notranslate">\; \&times;</span>{\mathrm{<span\; class="notranslate">\; 10</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; I</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; if</span>}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; \&Greater\; Equal;</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; I</span>}}\\ \mathrm{<span\; class="notranslate">\; 1.09</span>}<span\; class="notranslate">\; \&times;</span>{\mathrm{<span\; class="notranslate">\; 10</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; I</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; if</span>}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; \&le;</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; I</span>}}\end{array}\right.$

$\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 177.3</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; if</span>}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 177.3</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; 0</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; if</span>}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 177.3</span>}<span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; 0</span>}\end{array}\right.$

in, is the normalized quality score of the i-th sub-service of the user in the n period, *∈{V,A,T} indicates the type of the sub-service, V stands for video, A stands for voice, and T stands for text. is the average delay of the audio data queue, is the average delay of the image data queue, is the average delay of the text data queue.

Among them, in the calculation of the quality score of the video sub-service, AD is the quality score decrement caused by the absolute delay, and MS is the quality score decrement caused by the asynchronous image and audio.

In the step (2), the weighted quality score of each sub-service of the user in the current period is calculated according to the normalized quality score of each sub-service in the current period; according to the sorting of the weighted quality scores of each sub-service in the current period, the ROC algorithm is used to calculate Incremental attention of each sub-service of the user in the current period.

The formula for calculating the weighted quality score is as follows:

${\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; if</span>}\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; \&le;</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; B</span>}\\ \mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&beta;</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; if\; B</span>}<span\; class="notranslate">\; \&le;</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.$

in: is the weighted quality score of the i-th sub-service of the user in the n period, is the normalized quality score of the i-th sub-service of the user in the n period; *∈{V,A,T} indicates the type of sub-service, β ^* indicates the relevant weight of the sub-service type, 0<β ^T <β ^A < β ^V <1; B is the normalized quality score threshold, that is, the lowest normalized quality score acceptable to meet user communication needs, 0<B<1.

The formula for calculating the user attention of each sub-service in the step (2) is as follows:

a _i (n)=γa _i (n-1)+(1-γ)Δa _i (n)

Among them: a _i (n) and a _i (n-1) are the user attention degree of the user to the i-th sub-service in the n period and n-1 period respectively, and γ is the user attention degree in the previous period and the user attention in the current period degree influence weight, 0≤γ≤1; Δa _i (n) is the increment of user's attention to the i-th sub-service in n period.

A device for dispatching instant message services with fair user experience, comprising:

The data queue delay information maintenance module is used to store the information of the data queues of each sub-service, calculate and output the average delay of the data queues;

The user experience estimation module is used to calculate and output the representative service quality score of the user in the current period according to the average delay;

The sub-service data flow scheduling module is used to select the user with the worst service quality score within the coverage of the base station in the current period, and schedule a sub-service of the user with the user's attention degree as the probability.

Described user experience estimation module comprises:

The normalized quality score calculation submodule is used to calculate and output the normalized quality score of each sub-service in the current period according to the average delay;

The user attention degree maintenance sub-module is used to calculate the weighted quality score of each sub-service of the user in the current period according to the normalized quality score of each sub-service in the current period; and then use the ROC algorithm to calculate the increase in the attention degree of each sub-service of the user in the current period ; According to the user attention degree of each sub-service in the previous period and the attention increment of each sub-service in the current period, calculate and output the user attention degree of each sub-service in the current period;

The user representative service quality score calculation sub-module selects the sub-service with the highest user concern in the current period as the user's representative service, and reads the corresponding normalized quality score from the normalized quality score calculation sub-module, so that the representative service The normalized quality score of is used as the user's representative service quality score and output.

The beneficial effects of the present invention are as follows: (1) The present invention can estimate user experience in a scenario where there are multiple instant messaging users in the network and each user uses one or more sub-services, and realize user experience fairness through data flow scheduling. (2) The dispatching method of the present invention makes a dispatching decision through calculation according to the characteristics of the instant message service itself and user behavior habits. Therefore, no user feedback is required, the implementation complexity is low, and the engineering practicability is strong. (3) Under the condition that the user end buffers the data queue and reports the delay information of the data queue, the present invention is also applicable to the scheduling of the uplink transmission of the instant message service. (4) The present invention is applicable to all network scenarios that have a star topology and adopt central control and scheduling.

Description of drawings

Fig. 1 is a schematic diagram of the technical solution of the present invention.

Fig. 2 is a flow chart of the instant message service dispatching method for user experience fairness in the present invention.

FIG. 3 is a structural diagram of an instant message service scheduling device for user experience fairness according to the present invention.

Fig. 4 is a performance simulation result diagram of the method of the present invention.

detailed description

In order to describe the present invention more specifically, the technical solutions and related principles of the present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

The present invention is applicable to the scheduling of downlink instant message service of the cellular cell. Users are connected in star form with base stations. All instant message service users compete for the current available bandwidth (that is, the remaining bandwidth after satisfying guaranteed service users). The base station acts as a central controller to allocate resources and implement centralized scheduling of all instant message service users. Only one data flow gets a chance to be transmitted in any one time period. Among the three sub-services of the instant message service, the video sub-service includes two independent data streams of an image stream and a voice stream, the voice sub-service includes a voice data stream, and the text sub-service includes a text data stream.

The technical solution of the present invention is based on the following facts: (1) All sub-services of the instant message service are real-time or quasi-real-time services. Due to the user's own physical limitations, the user can only pay attention to one sub-service at any time period. (2) Users usually have a tendency to use the instant message sub-service, and the order of preference is video, voice, and text. (3) The degree of user attention to a sub-service changes with the quality of the sub-service, which will increase the user's attention to the sub-service with good quality and reduce the user's attention to the sub-service with poor quality.

The technical solution of the present invention is based on the delay analysis of the user sub-service data queue, tracks the user sub-service quality, estimates the user's attention to the sub-service, determines the user representative service, and realizes User experience is fair.

The principle of the technical solution of the present invention is as follows: (1) each sub-service is paid attention to by its users with a certain probability, and this probability is called the user attention degree of the sub-service, and the sum of the user attention degrees of all sub-services of a user at any time period is 1. In the initial period, the user attention degree of each sub-service of a user is obtained using the Randed Order Centroid (ROC) algorithm, which is all expectations that meet the ranking value in a statistical sense. The ordering of the sub-services required by the ROC algorithm is obtained according to the user's tendency to use the sub-services, that is, the video is in the front, the voice is in the middle, and the text is in the end. (2) The user attention degree of a sub-service changes with the quality of the sub-service. When the quality is good, the user attention degree increases, and when the quality is poor, the user attention degree decreases. The current user attention degree is the weighted sum of the user attention degree in the previous period and the current attention degree increment (3) The sub-service with the highest user attention degree is called the user representative service of the user in this period, and its quality is called the user representative service quality , indicating the user experience, and the higher the quality score, the better the experience. (4) The scheduling method of the present invention assigns the transmission opportunity to the subservice of the user (called the user to be served) with the lowest service quality score (that is, the worst user experience) on behalf of the current user, so that the user experience of all users tends to be similar. (5) The scheduling method of the present invention gives a transmission opportunity to a sub-service of the user to be served. The sub-service is selected according to probability from all the sub-services of the user, and the probability of a sub-service being selected is equal to the user attention of the sub-service. Spend. Probabilistic selection of sub-services not only maximizes the possibility of user representative services getting transmission opportunities, but also avoids the continuous impact caused by accidental estimation errors of user attention.

The instant message service dispatching method of user experience fairness of the present invention, as shown in Figure 1, comprises the following steps:

A. Data queue delay information maintenance;

B. User Experience Estimation;

C. User experience fair scheduling based on user representative service quality.

In the step A, the base station maintains its delay information for each data queue. The update of the average delay of the data queue adopts the sliding window method. The base station records the delay of the first data packet in the queue and the delay of the previous p-1 data packets, and performs arithmetic mean to obtain the average delay of the data queue. The base station maintains this delay information and updates this information every time the system sends a data packet.

In the step B, the current user attention degree of the sub-service is obtained through recursive calculation, and its value is the weighted sum of the user attention degree in the previous period and the current attention degree increment. The initial user attention is sorted according to the user's tendency to use the sub-service types, and is obtained by using the ROC algorithm. The preference ranking is based on user behavior habits and types of sub-services used by users. The attention increment is calculated by the ROC algorithm based on the ranking of the weighted quality scores of all sub-services. The weighted quality score of a sub-service is the sum of the part of the normalized quality score of the sub-service that does not exceed the threshold and the part that exceeds the threshold weighted by the weight of the sub-service. The sub-service normalized quality score is based on the average delay information of the data queue obtained in step A, combined with the sub-service type to which each data queue belongs (the video sub-service contains two data queues), and the quality scoring method of each sub-service is used for scoring , and normalized by its highest attainable quality score. The sub-service with the highest attention of each user is selected as the user's representative service, and the normalized quality score of the user's representative service is used as a measure of user experience.

In the step C, according to the user representative service quality score obtained in the step B, the user with the lowest score is selected as the user to be served, and the data flow of a sub-service is selected for scheduling in all current sub-services with the probability of user attention. If the selected sub-service contains 2 data streams (video sub-service), select the data stream with the largest average delay for scheduling.

A device for dispatching instant message services with fair user experience, comprising:

The data queue delay information maintenance module is used to store the information of the data queues of each sub-service, calculate and output the average delay of the data queues;

The user experience estimation module is used to calculate and output the representative service quality score of the user in the current period according to the average delay;

The sub-service data flow scheduling module is used to select the user with the worst service quality score within the coverage of the base station in the current period, and schedule a sub-service of the user with the user's attention degree as the probability.

The User Experience Estimation module includes:

The normalized quality score calculation submodule is used to calculate and output the normalized quality score of each sub-service in the current period according to the average delay;

The user attention degree maintenance sub-module is used to calculate the weighted quality score of each sub-service of the user in the current period according to the normalized quality score of each sub-service in the current period; and then use the ROC algorithm to calculate the increase in the attention degree of each sub-service of the user in the current period ; According to the user attention degree of each sub-service in the previous period and the attention increment of each sub-service in the current period, calculate and output the user attention degree of each sub-service in the current period;

The user representative service quality score calculation sub-module selects the sub-service with the highest user concern in the current period as the user's representative service, and reads the corresponding normalized quality score from the normalized quality score calculation sub-module, so that the representative service The normalized quality score of is used as the user's representative service quality score and output.

Embodiment one

As shown in Figure 2, the instant message service dispatching method of the present invention user experience fairness comprises the following steps:

Step 21: Initialize

Reset the packet counters for each data queue to zero. Set the average queuing delay record value of each data queue to zero. Calculate the initial user attention according to Table 1. In Table 1, the video service user is recorded as user V; the voice service user is recorded as user A; the video text user is recorded as user V+jT; the voice text user is recorded as user A+jT; the text user is recorded as user jT. Among them, j is the number of text sub-services used by users. In Table 1, α is the symbol of user attention; 0 means period 0, which is the initial period; the superscript indicates the type of sub-service, V represents video, A represents voice, and T represents text; subscript l represents the lth sub-service of the user, l =1,2,3...j. In the table, each user only uses one video or voice sub-service, so the subscript of the user attention degree of the video or voice sub-service is 1.

Table 1 Calculation method of initial user attention

Step 22: Data queue delay information maintenance

The base station maintains its delay information for each data queue, and updates the average delay of the data queue every time a data packet is successfully transmitted in the system

${\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\underset{\mathrm{<span\; class="notranslate">\; w</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; q</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; d</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; w</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; if\; q</span>}<span\; class="notranslate">\; \&Greater\; Equal;</span>\mathrm{<span\; class="notranslate">\; p</span>}\\ \underset{\mathrm{<span\; class="notranslate">\; w</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; q</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; d</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; w</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; otherwise</span>}\end{array}\right.<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>$

in: is the average delay of the data queue when the qth data packet in the data queue becomes the head of the queue, and d ^* (w) is the waiting delay of the wth data packet in the data queue, which is recorded by the base station. p is the length of the sliding window, and *∈{V,A,T} represents the type of sub-service, which respectively represent the video sub-service V, the voice sub-service A and the text sub-service T.

Step 23: Sub-business normalized quality score calculation

Sub-business normalized quality score According to the average delay of the data queue obtained in step 22, the sub-service quality scoring method can be used to score, and normalized by the highest attainable quality score. The various sub-service scoring methods and the highest attainable quality scores are shown in Table 2.

Table 2 Scoring method for sub-service quality of instant messaging service

In the calculation of the video sub-service quality score, AD is the decrement of the quality score brought by the absolute delay, MS is the decrement of the quality score caused by the asynchronous image and audio, $\mathrm{MS}=\left\{\begin{array}{cc}-1.38\&times;{10}^{-3}({\tilde{d}}_S-{\tilde{d}}_I),& \mathrm{if}{\tilde{d}}_S\&Greater\; Equal;{\tilde{d}}_I\\ 1.09\&times;{10}^{-3}({\tilde{d}}_S-{\tilde{d}}_I),& \mathrm{if}{\tilde{d}}_S\&le;{\tilde{d}}_I\end{array}\right.;$ is the average delay of the audio data queue, in ms, is the average delay of the image data queue, in ms. In the voice sub-service quality score calculation, the sign function $h({\tilde{d}}_S-177.3)=\left\{\begin{array}{cc}0& \mathrm{if}{\tilde{d}}_S-177.3<0\\ 1& \mathrm{if}{\tilde{d}}_S-177.3>0\end{array}\right.,$ It is the average delay of the audio data queue, in ms. In the text subservice quality score calculation, is the average delay of the text data queue, in ms.

Step 24: Sub-business weighted quality score update

The weighted quality of sub-services is the user's subjective evaluation of the sub-services used, covering the quality of the sub-service itself and the user's tendency to use such services, and is used for comparison among all sub-services of a user. According to the normalized quality score of each sub-service obtained in step 23, the weighted quality score of the sub-service in the current period can be calculated by formula (2).

${\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; if</span>}\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; \&le;</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; B</span>}\\ \mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&beta;</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; )</span>& \mathrm{<span\; class="notranslate">\; if\; B</span>}<span\; class="notranslate">\; \&le;</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

in: is the weighted quality score of the i-th sub-service of the user in the n period, is the normalized quality score of the i-th sub-service of the user in the n period; *∈{V,A,T} indicates the type of sub-service, β ^* indicates the relevant weight of the sub-service type, 0<β ^T <β ^A < β ^V <1; B is the normalized quality score threshold, that is, the lowest normalized quality score acceptable to meet user communication needs, 0<B<1.

Step 25: User Attention Update

The user attention degree of the i-th sub-service of the user in the current period is the weighted sum of the user attention degree in the previous period and the attention degree increment in the current period, and is calculated by formula (3).

a _i (n)=γa _i (n-1)+(1-γ)Δa _i (n) (3)

Among them: a _i (n) and a _i (n-1) are the user attention degree of the i-th sub-service by the user in the n period and n-1 period respectively, and γ is the user attention degree in the n-1 period to the user in the n period The influence weight of the attention degree, 0≤γ≤1; Δa _i (n) is the increment of the user's attention degree to the i-th sub-service in the n period, which can be positive or negative.

According to the weighted quality score of the sub-services obtained in step 24, the attention increment Δa _i (n) of the i-th sub-service of the user in the n period can be calculated by formula (4).

$\mathrm{<span\; class="notranslate">\; \&Delta;</span>}{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; |</span>}\underset{\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>}{\overset{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; |</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; r</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

Among them: m _i (n) is the ranking of the weighted quality scores of the i-th sub-service of the user in n period, |S _k | is the number of sub-services used by the user.

Step 26: User Representative Service Quality Score Calculation

The user representative service quality score is the normalized quality score of the sub-service with the highest user attention. QoS The score of can be obtained by formula (5):

$\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>\stackrel{\mathrm{<span\; class="notranslate">\; \&Delta;</span>}}{<span\; class="notranslate">\; =</span>}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; arg</span>}\underset{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; \&le;</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; |</span>}{\mathrm{<span\; class="notranslate">\; max</span>}}{\mathrm{<span\; class="notranslate">\; a</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

in: and a _i (n) are the normalized quality score and user attention of the i-th sub-service of the user in the n period, respectively, and i' is the sequence number of the sub-service with the highest user attention in the n period.

Step 27: Sub-service data flow scheduling

Based on the user representative service quality score obtained in step 26, first select the user with the lowest user representative service quality score as the user k _sel (n) to be served, as shown in formula (6)

${\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; sel</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; arg</span>}\underset{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; K</span>}}{\mathrm{<span\; class="notranslate">\; min</span>}}\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

Wherein, k represents the serial number of the user, and the total number of users within the coverage of the base station is K.

The serial number of the sub-service for which the selected user obtains the transmission opportunity is i _sel (n), which is selected according to formula (7).

$\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; i</span>}}_{\mathrm{<span\; class="notranslate">\; sel</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; a</span>}}_{{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; sel</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; <</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; i</span>}}_{\mathrm{<span\; class="notranslate">\; sel</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; a</span>}}_{{\mathrm{<span\; class="notranslate">\; k</span>}}_{\mathrm{<span\; class="notranslate">\; sel</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

Among them: h is randomly generated by the algorithm, and the value is between 0 and 1. If the selected i-th sub-service contains 2 data streams (video-type sub-service), select the data stream with the largest average delay for scheduling.

It can be seen from the above steps that the present invention records the average delay of all sub-service data queues of the user, uses a low-complexity method to estimate the user attention of the sub-service, determines the user representative service, and obtains the user representative service quality score as the user experience measure. User experience can be easily estimated without user feedback, and experience fairness among instant messaging users can be achieved by improving the worst user experience.

Embodiment two

The method for dispatching instant messages with fair user experience provided by the present invention can be realized by using the device shown in FIG. 3 . Devices include:

The data queue delay information maintenance module inputs the scheduled instant message data queue information, including the data flow type of each queue, the sub-services it belongs to and the users it belongs to. Monitor the delay of the sub-service data queue, update the delay data in the sliding window, calculate and update and output the average delay data of the data queue;

The user experience estimation module reads the average delay data of the data queue from the data queue delay information maintenance module, estimates and outputs the user representative service quality scoring data of all current instant message users;

The sub-service data flow scheduling module reads the user representative service quality scoring data of instant message users from the user experience estimation module, and outputs scheduling instructions.

The user experience estimation module further includes:

The normalized quality score calculation sub-module reads the average delay data of the data queue from the data queue delay information maintenance module, calculates and outputs the sub-service normalized quality score data;

The user attention degree maintenance sub-module reads the normalized quality score data of the sub-services from the normalized quality score calculation sub-module, updates the weighted quality scores of the sub-services, and updates and outputs the user attention degree data;

User representative service quality score calculation sub-module, reads sub-service normalized quality score data from the normalized quality score calculation sub-module, and reads user attention data from the user attention maintenance sub-module, calculates updates and outputs instant messages The user of the user represents the service quality scoring data.

This device has strong independence and does not require user participation. It can be added to the network dispatching system as an independent module without any modification to the original system.

The beneficial effects of the technical solution of the present invention can be verified through simulation. For example, in the following simple scenario: the number of users in the network is 20, of which 4 users (20%) use the video sub-service (can use the text sub-service at the same time), and 6 (20%) use the voice sub-service (can use the text sub-service at the same time). 30%), and the remaining 10 users only use the text sub-service. The number of text sub-services used by each user is 1 to 5, determined randomly. γ=0.7, Normalized quality score threshold B=0.6. As shown in Figure 4, compared with the traditional Round Robin (RR) scheduling and Minimum Quality First (MQF) scheduling, in the bandwidth range of 120-840kbps, the fairness of the scheduling method of the present invention (MRSQF) The indicators are all good.

Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and equivalent technologies, the present invention also intends to include these modifications and variations.

Claims (4)

1. A method for dispatching instant message services with fair user experience, comprising the steps of: (1) According to the average delay of the sub-service data queue, calculate the normalized quality score of each sub-service operated by the user in the current period according to the following formula; ${\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; A</span>}\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \}</span><span\; class="notranslate">\; ,</span>& <span\; class="notranslate">\; *</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; V</span>}\\ \mathrm{<span\; class="notranslate">\; 93.2</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.024</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 0.11</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 177.3</span>}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 177.3</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& <span\; class="notranslate">\; *</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; A</span>}\\ \mathrm{<span\; class="notranslate">\; 4.73</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; ln</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; 10</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3</span>}}{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; T</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& <span\; class="notranslate">\; *</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; T</span>}\end{array}\right.$ AD＝3.92-3.24×10 ^-4 (d _S +d _I ) $\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1.38</span>}<span\; class="notranslate">\; \&times;</span>{\mathrm{<span\; class="notranslate">\; 10</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; I</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; f</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; \&Greater\; Equal;</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; I</span>}}\\ \mathrm{<span\; class="notranslate">\; 1.09</span>}<span\; class="notranslate">\; \&times;</span>{\mathrm{<span\; class="notranslate">\; 10</span>}}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; I</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; f</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; \&le;</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; I</span>}}\end{array}\right.$ $\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 177.3</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; f</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 177.3</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; 0</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; f</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; S</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 177.3</span>}<span\; class="notranslate">\; ></span>\mathrm{<span\; class="notranslate">\; 0</span>}\end{array}\right.$ in, is the normalized quality score of the i-th sub-service of the user in the n period, *∈{V, A, T} indicates the type of the sub-service, V stands for video, A stands for voice, T stands for text, d _S is the number of audio data queues Average time delay, d _I is the average time delay of the image data queue, and d _T is the average time delay of the text data queue; (2) According to the normalized quality score of each sub-service in the current period, the weighted quality score of each sub-service of the user in the current period is calculated by the following formula, and then according to the ranking of the weighted quality scores of each sub-service in the current period, the ROC algorithm is used to calculate the current Incremental attention of each sub-service in the time period; calculate the user attention of each sub-service in the current period according to the user attention of each sub-service in the previous period and the increase in the attention of each sub-service in the current period; ${\mathrm{<span\; class="notranslate">\; Q</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; +</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; f</span>}\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; \&le;</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; B</span>}\\ \mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&beta;</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>& \mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; f</span>}\mathrm{<span\; class="notranslate">\; B</span>}<span\; class="notranslate">\; \&le;</span>{\stackrel{<span\; class="notranslate">\; ~</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right.$ Among them: Q ⁺ _i (n) is the weighted quality score of the i-th sub-service of the user in the n period, is the normalized quality score of the i-th sub-service of the user in the n period, β ^* represents the relevant weight of the sub-service type; B is the normalized quality score threshold; The ROC algorithm is the sorting center of gravity algorithm, and the sorting adopted by the algorithm is obtained according to the user's tendency to use the sub-services, that is, the video is in the front, the voice is in the middle, and the text is at the end; (3) Select the sub-service with the highest user concern in the current period as the representative service of the user, and make the normalized quality score of the representative service be the representative service quality score of the user; (4) Traverse all users within the coverage of the base station according to steps (1)-(3), get the representative service quality score of each user, select the user with the worst representative service quality score as the user to be served, and take the user attention as Probabilistically select a sub-service of the user to be served for service, so as to improve the normalized quality score of the sub-service of the user to be served. 2. The instant message service scheduling method according to claim 1, characterized in that: when the sub-service of the selected user to be served is served in the step (4), if the sub-service includes two data streams, then Provide services for data streams with large average delays. 3. the instant message service dispatching method according to claim 1, is characterized in that: the calculation formula of the degree of user attention of each sub-service in the described step (2) is as follows: a _i (n)=γa _i (n-1)+(1-γ)Δa _i (n) Among them: a _i (n) and a _i (n-1) are the user attention degree of the user to the i-th sub-service in the n period and n-1 period respectively, and γ is the user attention degree in the previous period and the user attention in the current period Δa _i (n) is the increment of user's attention to the i-th sub-service in the n period. 4. A device for dispatching instant message services with fair user experience, characterized in that it comprises: The data queue delay information maintenance module is used to store the information of the data queues of each sub-service, calculate and output the average delay of the data queues; The user experience estimation module is used to calculate and output the representative service quality score of the user in the current period according to the average delay; The sub-service data flow scheduling module is used to select the user with the worst service quality score within the coverage of the base station in the current period, and schedule a sub-service of the user with the probability of user attention; Described user experience estimation module comprises: The normalized quality score calculation submodule is used to calculate and output the normalized quality score of each sub-service in the current period according to the average delay; The user attention degree maintenance sub-module is used to calculate the weighted quality score of each sub-service of the user in the current period according to the normalized quality score of each sub-service in the current period; and then use the ROC algorithm to calculate the increase in the attention degree of each sub-service of the user in the current period ; According to the user attention degree of each sub-service in the previous period and the attention increment of each sub-service in the current period, calculate and output the user attention degree of each sub-service in the current period; The user representative service quality score calculation sub-module selects the sub-service with the highest user concern in the current period as the user's representative service, and reads the corresponding normalized quality score from the normalized quality score calculation sub-module, so that the representative service The normalized quality score of is used as the user's representative service quality score and output; The ROC algorithm is the sorting center of gravity algorithm, and the sorting adopted by the algorithm is obtained according to the user's tendency to use the sub-services, that is, the video is in the front, the voice is in the middle, and the text is in the last.