Resource preemption method based on maximization of user satisfaction

Abstract

The invention provides a resource preemption method based on user satisfaction maximization, and belongs to the technical field of resource allocation. On the basis of a centralized network architecture, a resource preemption cost function model is established according to four optimization standards of preemption service priority, preemption service quantity, preemption bandwidth and user satisfaction, then all preemption services are arranged in an ascending order according to cost values, different schemes are executed according to different configurations of weights when the cost values are the same, and the preemption cost function model is optimized. And finally seizing the services in sequence. By designing a set of bandwidth resource preemption strategy, the problem of transmission of high-priority burst service requests can be solved, and the influence on the service to be preempted is reduced as far as possible while the important service delivery rate and the user satisfaction degree are improved, so that the system transmission value is improved.

Description

Technical Field

[0001] The invention relates to the technical field of resource allocation, and in particular to a resource preemption method based on maximizing user satisfaction. Background Art

[0002] In recent years, with the rapid development of mobile communication technology, cloud computing technology and Internet applications, a large number of mobile terminals and intelligent Internet of Things devices are constantly connected to the network, resulting in the switching system in the communication network experiencing an explosive growth in network service traffic. At the same time, emerging services have also put forward different QoS transmission requirements for the Internet. The transmission capacity of each link in the network is different. Reasonable scheduling of bandwidth resources for business flows is a prerequisite for efficient and reliable information transmission and effective utilization of network resources. Therefore, it is necessary to design a reasonable resource scheduling scheme based on the current network status and different business requirements. Due to the suddenness and unpredictability of business requests, suboptimal decisions often occur in traditional routing planning. Therefore, the present invention designs a resource preemption scheme to achieve bandwidth resource scheduling. Summary of the invention

[0003] In view of this, the present invention proposes a resource preemption method based on maximizing user satisfaction based on a centralized network architecture. By designing a set of bandwidth preemption strategies, the transmission problem of high-priority burst service requests is solved, the delivery rate of important services is improved while minimizing the impact on the preempted services, thereby improving the transmission value of the system.

[0004] Based on the above purpose, the technical solution provided by the present invention is:

[0005] A resource preemption method based on maximizing user satisfaction includes the following steps:

[0006] Step 1: Establish a resource preemption cost function model based on four optimization criteria: preemption service priority, preemption service quantity, preemption bandwidth, and user satisfaction. The resource preemption cost function model is:

[0007]

[0008] Where H(l) represents the cost of preempting service l; α, β and γ are the weights of the three objective subfunctions; l represents the lth element in the set L, which is the set of preemptible services to be selected. The priority of maintaining services in the set is lower than the priority of establishing new services.

[0009] It refers to the value cost introduced by the amount of data transmitted by the service when preempting service l. K(l) represents the priority cost of the preempted service l. K(l) = 8-p(l), where p(l) is the retention priority of service l. w represents the completion degree of service transmission, w∈(0,1);

[0010] It refers to the cost of reducing the number of preempted connections when preempting service l, and b(l) is the bandwidth occupied by service l;

[0011] γ(b(l)-r) 2 It refers to the cost of bandwidth waste caused by preempting service l, r is the bandwidth that needs to be preempted by new or rerouted service ρ, and service ρ is the bandwidth of transmission link e. j The bandwidth required to be seized is r = b(ρ)-a j , a j It is the transmission link j The remaining available bandwidth is b(ρ), and b(ρ) is the bandwidth requirement of service ρ;

[0012] Step 2: sort all preemptible services in ascending order according to the cost value H(l); and determine whether there are the same cost value H(l). If so, execute different schemes through different configurations of α, β, and γ to obtain the sorted service set L';

[0013] Step 3: According to the sorted service set L' obtained in step 2, preempt the services in L' to obtain the preempted service set S and the wasted bandwidth Waste;

[0014] Step 4: Eliminate the services with bandwidth smaller than Waste in the preempted service set S to obtain the final preempted service set.

[0015] Furthermore, the specific process of step 2 is:

[0016] Arrange all preemptible services in L in ascending order according to the cost value H(l), and determine whether there are the same cost value H(l). If so, execute different solutions through different configurations of α, β and γ:

[0017] If the α value is the largest, the services are sorted in ascending order according to the value v(l); if v(l) are the same, they are sorted according to (b(l)-r) 2 Sort the business in ascending order, if (b(l)-r) 2 If they are the same, the services are sorted in descending order according to b(l);

[0018] If the β value is the largest or the three are equal, then sort the services in descending order according to b(l). If b(l) are the same, then sort the services in ascending order according to the value v(l). If v(l) are the same, then sort the services in ascending order according to (b(l)-r). 2 Sort the businesses in ascending order;

[0019] If the γ value is the largest, then (b(l)-r) 2 Sort the business in ascending order, if (b(l)-r) 2If they are the same, the services are sorted in ascending order according to the value cost v(l). If v(l) is the same, the services are sorted in descending order according to b(l).

[0020] Furthermore, the specific process of step 3 is as follows:

[0021] Step 301: Preempt the first service in L'.

[0022] Step 302: Determine whether the preempted bandwidth Preempt of the current link is less than the bandwidth r to be preempted.

[0023] Step 303: If Preempt < r, deposit the first service l' in L' into the set S, delete the corresponding service in the set L', let Preempt = Preempt + b(l'), and return to step 301; otherwise, calculate the wasted bandwidth Waste = Preempt - r to obtain the preempted service set S and the wasted bandwidth Waste; where b(l') is the bandwidth occupied by the service l'.

[0024] Furthermore, the specific process of step 4 is as follows:

[0025] Step 401: Determine whether there is a service in S with a bandwidth less than or equal to Waste.

[0026] Step 402: If there is a service in S with a bandwidth less than or equal to Waste, sort the services that meet the requirements in descending order according to the loss cost value to obtain the set W'. Remove the first service s', and let Waste = Waste - b(s'), where b(s') is the bandwidth occupied by the service s', and return to step 401; otherwise, obtain the final preempted service set.

[0027] As can be seen from the above description, the beneficial effects of the technical solution of the present invention are as follows:

[0028] In an actual network, administrators rarely care about improving user satisfaction. In a QoS network, users consider service continuity to be the most important. When preemption is inevitable, the continuity of the preempted service will be interrupted. If the service is preempted just when it starts to be transmitted, the reduction in user satisfaction is relatively small. Otherwise, it will cause a higher degree of dissatisfaction. This solution reduces the probability of the service flow about to be transmitted being preempted by introducing the concept of service transmission completion degree into the cost function. Therefore, this solution can effectively improve user satisfaction. Description of the Drawings

[0029] Figure 1 It is a flowchart of the resource preemption scheme according to the embodiment of the present invention.

[0030] Figure 2This is a comparison chart between the solution of the present invention and the traditional v-prept solution in terms of waste of preempted service bandwidth.

[0031] Figure 3 The figure is a comparison chart of the average priority of preempting services between the solution of the present invention and the traditional v-prept solution.

[0032] Figure 4 The figure is a comparison chart of the average transmission completion degree of the preempted service between the solution of the present invention and the traditional v-prept solution. DETAILED DESCRIPTION

[0033] In order to facilitate the understanding of the technical solution of the present invention by those skilled in the art, and at the same time, to make the technical purpose, technical solution and beneficial effects of the present invention clearer and to fully support the scope of protection of the claims, the technical solution of the present invention is further and more detailedly explained in the form of specific cases below.

[0034] An intelligent routing method based on maximizing user satisfaction. In this method, values ​​are first set for different types of services, and service characteristics, service value, and topology information are input into a deep neural network structure. The neural network outputs the next hop node of the path, and then the node is used as the source node by the neural network to continue to output the next hop. After the complete path is output, the reward is calculated, and the network is trained according to feedback until the system converges.

[0035] The specific steps include:

[0036] Step 1: Establish a resource preemption cost function model, specifically:

[0037] The resource preemption cost function model is constructed based on the following four optimization criteria:

[0038] 1) Preempt lower priority traffic to avoid cascading preemption;

[0039] 2) Seize the least number of business flows and reduce the number of affected businesses;

[0040] 3) Minimize bandwidth grabbing and avoid bandwidth waste;

[0041] 4) Improve user satisfaction and avoid preemption of business flows that are about to be transmitted.

[0042] Based on this, the resource preemption cost function model H(l) is:

[0043]

[0044] in:

[0045] α, β, and γ are the weights of the three objective sub-functions. By configuring α, β, and γ, different treatments can be achieved for each optimization standard, and the execution order of the optimization standards can be flexibly changed according to the specific configuration of the weights by the network administrator, and preemption is performed according to this order. If the α value is the largest, the goal is to optimize the preemption priority; if the β value is the largest, the goal is to optimize the number of preempted services; if the γ value is the largest, the goal is to optimize the preempted bandwidth; if the α, β, and γ values ​​are equal, the optimization goal is to optimize the preempted bandwidth; the default goal is to optimize the preempted bandwidth.

[0046] l represents the lth element in the set L;

[0047] L is a set of preemptible services to be selected. The maintenance priority of the services in the set is lower than the establishment priority of the new service. There are N elements in the set.

[0048] K(l) represents the lth element in the priority cost set K, which is the priority cost of the preempted service l;

[0049] K is a set of priority costs, which can be represented by K(1), K(2), ..., K(l), ..., K(N). When calculating it, it is generally believed that the preemption cost K(l) is linearly related to the retention cost p(l), and the K(l) value is calculated using the formula K(l) = 8-p(l), where p(l) is the retention priority of service l.

[0050] It refers to the value cost introduced by the amount of data transmitted by the service l when preempting the service l. K(l) = 8-p(l) represents the service value and is related to the service priority. Since service continuity will be interrupted when preemption occurs, the larger the amount of data transmitted, the higher the cost of loss caused by preemption, and the lower the user satisfaction.

[0051] It refers to the cost of reducing the number of preempted connections when preempting service l, and b(l) is the bandwidth occupied by service l.

[0052] γ(b(l)-r) 2 It refers to the cost of bandwidth waste caused by preempting service l, r is the bandwidth that needs to be preempted by new or rerouted service ρ, and service ρ is the bandwidth of transmission link e. j The bandwidth that needs to be seized r = b(ρ)-a j , a j It is the transmission link j The remaining available bandwidth is b(ρ), and b(ρ) is the bandwidth requirement of service ρ;

[0053] H(l) represents the total cost of preempting service l. The smaller H(l), the easier it is for service l to be preempted.

[0054] Step 2: sort all preemptible services in ascending order according to the cost value H(l); and determine whether there are the same cost value H(l). If so, different schemes are executed through different configurations of α, β, and γ to obtain the sorted service set L'; the specific process is:

[0055] Arrange all preemptible services in L in ascending order according to the cost value H(l), and determine whether there are the same cost value H(l). If so, execute different solutions through different configurations of α, β and γ:

[0056] If the α value is the largest, the services are sorted in ascending order according to the value v(l); if v(l) are the same, they are sorted according to (b(l)-r) 2 Sort the business in ascending order, if (b(l)-r) 2 If they are the same, the services are sorted in descending order according to b(l);

[0057] If the β value is the largest or the three are equal, then sort the services in descending order according to b(l). If b(l) are the same, then sort the services in ascending order according to the value v(l). If v(l) are the same, then sort the services in ascending order according to (b(l)-r). 2 Sort the businesses in ascending order;

[0058] If the γ value is the largest, then (b(l)-r) 2 Sort the business in ascending order, if (b(l)-r) 2 If they are the same, the services are sorted in ascending order according to the value cost v(l); if v(l) are the same, the services are sorted in descending order according to b(l).

[0059] Step 3: According to the sorted service set L' obtained in step 2, the services in L' are preempted to obtain the preempted service set S and the wasted bandwidth Waste; the specific process is as follows:

[0060] Step 301: Seize the first service in L';

[0061] Step 302: Determine whether the preempted bandwidth Preempt of the current link is less than the bandwidth r that needs to be preempted;

[0062] Step 303: If Preempt < r, then deposit the first service l' in L' into set S, delete the corresponding service from set L', let Preempt = Preempt + b(l'), and return to Step 301; otherwise, calculate the wasted bandwidth Waste = Preempt - r to obtain the preempted service set S and the wasted bandwidth Waste; where b(l') is the bandwidth occupied by service l'.

[0063] Step 4. Eliminate the services in the preempted service set S with a bandwidth less than Waste to obtain the final preempted service set.

[0064] The specific process is as follows:

[0065] Step 401: Determine whether there is a service in S with a bandwidth less than or equal to Waste;

[0066] Step 402: If there is a service in S with a bandwidth less than or equal to Waste, then sort the services that meet the requirements in descending order of loss cost value to obtain set W', eliminate the first service s', and let Waste = Waste - b(s'), where b(s') is the bandwidth occupied by service s', and return to Step 401; otherwise, obtain the final preempted service set.

[0067] Figure 2-Figure 4 This is a simulation comparison chart of the solution proposed by the present invention and the traditional v-prept solution in terms of wasted bandwidth of preempted services, average priority, and average transmission completion rate. The simulation uses MATLAB2017. Figure 2-Figure 4 The abscissa is the number of simulation times, totaling 1000 times. For each simulation, the service flows and new service flows transmitted in the link are randomly generated, the preempted service sets required by each solution are output, and the wasted bandwidth, average priority, and average transmission completion rate are calculated.

[0068] As can be seen from the figure, compared with the v-prept solution, although the proposed solution of the present invention has relatively more wasted bandwidth, the average priority and average transmission completion rate of the preempted services are less than those of the traditional solution. Therefore, it can be seen that the solution proposed by the present invention can significantly improve the delivery rate of important services and effectively improve user satisfaction.

[0069] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A resource preemption method based on maximizing user satisfaction, characterized in that: It includes the following steps: Step 1: Establish a resource preemption cost function model according to four optimization criteria, namely preemption service priority, number of preemption services, preemption bandwidth, and user satisfaction. The resource preemption cost function model is as follows: In the formula, H(l) represents the cost value required for preemption service l; α, β, and γ are the weights of three target sub-functions; l represents the l-th element in set L, and L is the set of candidate services for preemption. The holding priorities of services in the set are all lower than the establishment priority of the new service; It refers to the value cost introduced by the amount of data transmitted by the service when preempting service l. K(l) represents the priority cost of the preempted service l. K(l) = 8-p(l), where p(l) is the retention priority of service l. w represents the completion degree of service transmission, w∈(0,1); It refers to the cost of reducing the number of preempted connections when preempting service l, and b(l) is the bandwidth occupied by service l; γ(b(l)-r) 2 It refers to the cost of bandwidth waste caused by preempting service l, r is the bandwidth that needs to be preempted by new or rerouted service ρ, and service ρ is the bandwidth of transmission link e. j The bandwidth that needs to be seized r = b(ρ)-a j , a j It is the transmission link j The remaining available bandwidth is b(ρ), and b(ρ) is the bandwidth requirement of service ρ; Step 2: Sort all preemption services in ascending order according to the cost value H(l); and determine whether there are the same cost values H(l). If so, execute different solutions through different configurations of α, β, and γ to obtain the sorted service set L′; Step 3: According to the sorted service set L′ obtained in Step 2, preempt the services in L′ to obtain the preemption service set S and the wasted bandwidth Waste; Step 4: Eliminate the services in the preemption service set S with bandwidth less than Waste to obtain the final preemption service set.

2. A resource preemption method based on maximizing user satisfaction according to claim 1, characterized in that: The specific process of Step 2 is as follows: Sort all preemption services in L in ascending order according to the cost value H(l), and determine whether there are the same cost values H(l). If so, execute different solutions through different configurations of α, β, and γ: If the α value is the largest, the services are sorted in ascending order according to the value v(l); if v(l) are the same, they are sorted according to (b(l)-r) 2 Sort the business in ascending order, if (b(l)-r) 2 If they are the same, the services are sorted in descending order according to b(l); If the β value is the largest or the three are equal, then sort the services in descending order according to b(l). If b(l) are the same, then sort the services in ascending order according to the value v(l). If v(l) are the same, then sort the services in ascending order according to (b(l)-r). 2 Sort the businesses in ascending order; If the γ value is the largest, then (b(l)-r) 2 Sort the business in ascending order, if (b(l)-r) 2 If they are the same, the services are sorted in ascending order according to the value cost v(l); if v(l) are the same, the services are sorted in descending order according to b(l).

3. A resource preemption method based on maximizing user satisfaction according to claim 2, characterized in that: The specific process of Step 3 is as follows: Step 301: Preempt the first service in L′; Step 302: Determine whether the already preempted bandwidth Preempt of the current link is less than the bandwidth r to be preempted; Step 303: If Preempt < r, store the first service l′ in L′ into set S, delete the corresponding service in set L′, let Preempt = Preempt + b(l′), and return to Step 301; Otherwise, calculate the wasted bandwidth Waste = Preempt - r to obtain the preemption service set S and the wasted bandwidth Waste; where, b(l′) is the bandwidth occupied by service l′.

4. A resource preemption method based on maximizing user satisfaction according to claim 3, characterized in that: The specific process of Step 4 is as follows: Step 401: Determine whether there is a service in S with bandwidth less than or equal to Waste; Step 402: If there is a service in S with bandwidth less than or equal to Waste, sort the services that meet the requirements in descending order according to the loss cost value to obtain set W′, eliminate the first service s′, and let Waste = Waste - b(s′), where b(s′) is the bandwidth occupied by service s′, and return to Step 401; Otherwise, obtain the final preemption service set.

Images