Multi-objective resource allocation method for multi-service qos oriented low earth orbit satellite network

Abstract

The application discloses a multi-target resource allocation method for a low-orbit satellite network for multi-service QoS. According to the QoS requirements of a low-orbit carrying network service, such as service delay and bandwidth, and in combination with service priority, corresponding key services are extracted, and key service weights are generated. According to the spatial area range corresponding to the low-orbit satellite carrying network service, a satellite carrying the service is selected, and the number of satellites requiring allocated resources is reduced. With the available resource time window of the satellite as a constraint condition, the key service weight and the service bandwidth are taken as the resource allocation targets of the selected key service, the service success matching probability is taken as the resource allocation target of the ordinary service other than the key service, a non-dominated sorting genetic algorithm is adopted, a normalized resource input mode and a multi-target fitness function are designed, and under the premise of as far as possible meeting the QoS of the key service of the low-orbit carrying network, the amount of services required to be calculated by the genetic algorithm and the number of allocated resources are reduced, and the purpose of lightweight allocation of satellite resources is achieved.

Description

Technical Field

[0001] This application relates to the technical field of network and information processing, and in particular to a method for multi-target resource allocation in low-Earth orbit satellite networks oriented towards multi-service QoS. Background Technology

[0002] Low-Earth orbit satellite bearer networks are situated between space access networks and core networks. They can carry inter-satellite multi-service transmission and interconnection. Their inter-satellite transmission resources have dynamic and time-varying characteristics. They are geared towards the differentiated service quality requirements of multiple services, but lack the ability to optimize and allocate network-based resources for multiple objectives.

[0003] Existing research has proposed several multi-objective resource allocation methods, but they mainly focus on the following aspects: different performance indicators of the overall network, such as spectral efficiency, energy efficiency, resource consumption, resource utilization, and service completion time; multi-objective allocation strategies proposed based on the different characteristics of service carriers; and joint optimization allocation of bandwidth and power in multi-beam EOR / LOR joint systems. However, these methods do not address resource allocation based on different service QoS requirements. Furthermore, when dealing with LOR satellite resource allocation, the numerous constraints and input conditions of existing methods may lead to the absence of an optimal solution or difficulty in algorithm convergence, hindering the expansion of satellite-based lightweight computing. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies in ensuring differentiated QoS for critical services when allocating massive amounts of low-Earth orbit (LEO) satellite resources in the future, this application provides a multi-objective resource allocation method for LEO satellite networks oriented towards multi-service QoS. This method extracts differentiated QoS requirements for critical services and distinguishes different types of critical services; proposes a lightweight carrier network spatial area range to reduce the number of satellite resources that need to be calculated; and proposes a multi-objective resource allocation method, designing a normalized resource input mode and a multi-objective fitness function. While satisfying the differentiated QoS of critical services in the LEO carrier network as much as possible, it achieves the goal of lightweight satellite resource allocation by reducing the amount of services and resources required for computation by the genetic algorithm.

[0005] Firstly, a multi-objective resource allocation method for low-Earth orbit satellite networks oriented towards multi-service QoS is provided, including:

[0006] The satellite control center obtains initial status information of the satellite bearer network, including satellite link availability time, total satellite bandwidth resources, and available bandwidth resources.

[0007] Based on the latency, bandwidth, and QoS requirements of key services in the low-Earth orbit bearer network, and combined with service priorities, corresponding key services are extracted, and key service weights are generated.

[0008] Based on the spatial area range corresponding to the low-Earth orbit satellite carrier network services, select satellites to carry the services, reduce the number of satellites that need to be allocated resources, and establish a multi-objective constrained network model for the low-Earth orbit satellite carrier network.

[0009] To meet the QoS requirements of different service levels, a multi-objective optimization function is generated for the aforementioned low-Earth orbit satellite bearer network model.

[0010] By using resource normalization methods, the multi-objective function of the business is converted into a corresponding fitness function;

[0011] The optimal solution is obtained by solving a multi-objective optimization problem oriented towards the fitness function.

[0012] Based on the resource allocation optimization results obtained from the algorithm, the corresponding satellite bandwidth resources are allocated to the applied services.

[0013] In conjunction with the first aspect, in some implementations of the first aspect, if the delay requested by the business... Less than the delay threshold The applied latency jitter Less than the delay jitter threshold The requested packet loss rate Less than the packet loss rate threshold If so, the service is classified as a highly reliable deterministic service and denoted as... If the service requests bandwidth Greater than the bandwidth threshold If so, the service is classified as a high-bandwidth service and denoted as... The remaining business is classified as ordinary business and is denoted as... .

[0014] In conjunction with the first aspect, in some implementations of the first aspect, the final weight of each business is denoted as... The highest priority corresponding to each type of matrix is ​​denoted as . The weighting coefficients for each type of business are set as follows: The value obtained by rounding down from 2 is represented as the business weight. .

[0015] In conjunction with the first aspect, in some implementations of the first aspect, the multi-objective optimization function is as follows:

[0016] High-reliability deterministic business resource allocation optimization function:

[0017] High-bandwidth service optimized resource allocation function:

[0018] Common business resource allocation function: ;

[0019] :business Requested bandwidth resource values;

[0020] : The business successfully allocated a resource identifier, if the business If bandwidth resources are successfully allocated, then ,otherwise .

[0021] In conjunction with the first aspect, in some implementations of the first aspect, the multi-objective fitness function of the low-Earth orbit carrier network is as follows:

[0022] Highly reliable deterministic business fitness function: , The sum of the business weights for the applied business;

[0023] High-bandwidth service fitness function: , This represents the total satellite bandwidth resources contained in the network.

[0024] Typical business fitness function: , This represents the total available window time for all satellites included in the network.

[0025] In conjunction with the first aspect, in some implementations of the first aspect, the constraints of the multi-objective network model are as follows:

[0026] A service can only allocate bandwidth resources within one time window of a specific inter-satellite link on a specific satellite at any given time; it cannot allocate inter-satellite link bandwidth resources simultaneously across multiple time windows.

[0027] Businesses must allocate resources within the visible time window of the link;

[0028] Prioritize satisfying the objective function of highly reliable deterministic services, followed by the objective function of high-bandwidth services, and lastly the objective function of ordinary services.

[0029] In conjunction with the first aspect, in some implementations of the first aspect, the fitness function of the low-orbit carrier network ,in , , This is the adjustment coefficient of the fitness function.

[0030] In conjunction with the first aspect, in certain implementations of the first aspect, the multi-objective optimization solution oriented towards the fitness function to obtain the optimal solution includes:

[0031] An improved NSGA-II genetic algorithm is used to perform multi-objective optimization based on the fitness function, and the Parto optimal solution is obtained.

[0032] In conjunction with the first aspect, in some implementations of the first aspect, the method includes:

[0033] Set the population size, randomly generate an initial population, and adapt the business parameter information to the population data format required by the algorithm.

[0034] The fitness value of each operational chromosome in the population is calculated based on the fitness function;

[0035] Update satellite available bandwidth resources and time window information;

[0036] Perform non-dominated sorting, calculate the crowding distance of the sorted population, and perform adaptive crossover and mutation operations;

[0037] Generate the next generation of population;

[0038] Determine if the current generation is the maximum generation. If not, recalculate the fitness value of each business chromosome. If it is, obtain the Pareto optimal solution set under the current generation and obtain the current resource allocation status.

[0039] Secondly, an electronic device is provided for performing the method described in any implementation of the first aspect.

[0040] Compared with the prior art, the solution provided in this application has at least the following beneficial technical effects:

[0041] (1) Through the following steps 1 to 6, the present invention transforms the multi-objective requirements of services into specific QoS requirements for key services, proposes a multi-objective fitness function that matches the key services, and achieves the purpose of lightweight allocation of satellite resources under the premise of satisfying the differentiated QoS of key services in the low-orbit bearer network.

[0042] (2) The present invention uses the following steps 2a~2c, the key service extraction method, to compare the QoS requirements such as service transmission delay and service bandwidth with the preset QoS threshold, so as to accurately divide the key services and narrow the scope of services to be calculated.

[0043] (3) The present invention selects satellites to carry services based on the spatial area range corresponding to the low-orbit satellite carrying network services through the following steps 3a~3c, thereby reducing the number of satellites that need to be allocated resources.

[0044] (4) The present invention optimizes satellite resource allocation by designing a normalized resource input mode and a multi-objective fitness function through the following steps 6a~6h. Attached Figure Description

[0045] Figure 1 A flowchart illustrating a multi-target resource optimization allocation method for differentiated QoS of services.

[0046] Figure 2 This is a flowchart of a multi-objective resource allocation algorithm based on genetic algorithms.

[0047] Figure 3 This is a diagram illustrating the classification of key business operations.

[0048] Figure 4 This is a schematic diagram of an example of a low-Earth orbit satellite topology.

[0049] Figure 5 A diagram illustrating business weights that were not successfully assigned.

[0050] Figure 6 This is a diagram illustrating unallocated service bandwidth.

[0051] Figure 7 This is a diagram illustrating unsuccessfully assigned business time. Detailed Implementation

[0052] The present application will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0053] like Figure 3 As shown, this method categorizes services into three types based on their characteristics and service requirements: highly reliable deterministic services, high-bandwidth services, and ordinary services. Highly reliable deterministic services require the satellite network to transmit services reliably within a deterministic timeframe. They have strict requirements on performance aspects such as latency, jitter, and transmission bandwidth. Resource allocation should minimize conflicts to ensure timely and accurate service delivery. These services are typically used in emergency communications, operational needs, and other applications with extremely high reliability and latency requirements. This type of service has the highest weight among the different service types. High-bandwidth services place high demands on the carrying capacity of the satellite network link. Resource allocation should provide sustainable high-bandwidth resources for this service. These services are typically used in high-traffic broadband applications such as ultra-high-definition video and large-capacity image processing. Due to their high bandwidth resource requirements, this type of service also has a high weight. Ordinary services do not have strict QoS requirements, but they do have certain requirements for successful access. It is necessary to minimize packet loss. In terms of resource allocation, it is necessary to ensure the access and communication time of ordinary services while satisfying other high-weight services. This mainly targets application scenarios involving file, data, and other fragmented services. These types of services have the lowest weight, but their access volume needs to be maximized to improve the overall network resource utilization and throughput. In this invention, the division of different services is not absolute but can be dynamically adjusted according to the set QoS thresholds.

[0054] like Figure 1 As shown in the figure, this application provides a method for multi-target resource allocation in a low-Earth orbit satellite network for multi-service QoS. Specific implementation examples include: setting up a satellite network simulation environment: setting up 16 low-Earth orbit satellite nodes, with a topology evenly distributed across four orbital planes; setting available links and available time windows for each satellite based on its topological position; setting the number of services to 64; and setting input parameters for each service, including its service priority, bandwidth requirements, and service duration. The genetic algorithm simulation parameters are also set: population size is 60, the number of generations is 100, and the gene mutation probability is 0.2.

[0055] Step 1: The satellite control center obtains the initial status information of the satellite bearer network, including satellite link availability time, total satellite bandwidth resources, and available bandwidth resources.

[0056] Step 2: Based on the QoS requirements and application characteristics of the services, divide the services into three categories: high reliability deterministic services, high bandwidth services, and ordinary services, and calculate the service weights.

[0057] Step 2a: Receive service requests and obtain service QoS requirements (including latency, latency jitter, bandwidth requirements, and packet loss rate) and service priority information, where the service priority is denoted as... Business Priority It is divided into 8 levels, with higher priority numbers indicating higher priority.

[0058] Step 2b: Delay threshold based on set business requirements Delay jitter threshold Bandwidth threshold and packet loss rate threshold Determine each business category.

[0059] If the business requests a delay Less than the delay threshold The applied latency jitter Less than the delay jitter threshold The requested packet loss rate Less than the packet loss rate threshold If so, the service is classified as a highly reliable deterministic service and denoted as... If the service requests bandwidth Greater than the bandwidth threshold If so, the service is classified as a high-bandwidth service and denoted as... The remaining business is classified as ordinary business and is denoted as... In this example, the highly reliable deterministic service after partitioning. The corresponding priority matrix is ​​[8 7], high bandwidth services The corresponding priority matrix is ​​[6 5 4], for ordinary services. The corresponding priority matrix is ​​[3 21].

[0060] Step 2c: Record the final weight of each business as... The highest priority corresponding to each type of matrix is ​​denoted as . The weighting coefficients for each type of business are set as follows: The value obtained by rounding down from 2 is: =4、 =3 and =1, then the business weight can be expressed as There are 64 business items. Taking three typical business items as examples: Business 1 has a priority of 8 and a weighting coefficient of 4, so its business weight is 32; Business 7 has a priority of 6 and a weighting coefficient of 3, so its business weight is 18; Business 8 has a priority of 3 and a weighting coefficient of 1, so its business weight is 3.

[0061] Step 3: Establish a multi-objective constrained network model for the low-Earth orbit satellite carrier network. In this example, the network model selects 16 satellites to form a rectangular topology, as shown below. Figure 4 As shown, its network parameter settings are as follows:

[0062] : A set of low-Earth orbit satellite nodes, each satellite is represented as .

[0063] The number of low-orbit satellite nodes is 16 in this embodiment.

[0064] : It is a collection of low-Earth orbit satellite network services, among which For a set of highly reliable deterministic services, For high-bandwidth service collection, It is a collection of ordinary business processes, and Each business is represented as .

[0065] The number of low-Earth orbit satellite network services is 64 in this embodiment;

[0066] :business Request bandwidth resource values.

[0067] :business Application priority.

[0068] :business The converted weights are calculated in step 2.

[0069] :business Request communication time.

[0070] Each satellite The set of inter-satellite link bandwidth resources.

[0071] Single satellite The number of available communication time windows, .

[0072] Single satellite The set of communicable time windows, such as Figure 4 As shown, satellite S1, located at the topological edge, can only communicate with S5 and S2. It contains a maximum of two time windows. S6 in the middle region of the topology can communicate with S2, S10, S5, and S7. It can contain up to four time windows.

[0073] : The business successfully allocated a resource identifier, if the business If bandwidth resources are successfully allocated, then ,otherwise .

[0074] Business For example, its application for bandwidth resources For 310Mbps, request communication time The priority of the service request is 430(s). The business weight is 8. It is 32. (Based on satellite) For example, its available resources are represented as follows: the first to fourth rows represent its inter-satellite link resources with its neighboring satellites in front, behind, left, and right, respectively (these are referred to as the row numbers). (Multiple windows), the first column is the communicateable time window. (Unit: seconds), the second column shows available bandwidth resources. (Unit: Mbps), Number of available communication time windows 2:

[0075]

[0076] Step 4: Generate a multi-objective optimization function for the above-mentioned low-orbit satellite bearer network model, taking into account the QoS requirements of different service levels.

[0077] Step 4a: Based on the QoS requirements of different service levels, the following multi-objective optimization function for the low-orbit bearer network is generated:

[0078] High-reliability deterministic business resource allocation optimization function:

[0079] The high-reliability deterministic business resource allocation optimization function represents the goal of successfully allocating this type of business by maximizing the weight of the allocated business.

[0080] High-bandwidth service optimized resource allocation function:

[0081] The high-bandwidth service optimization resource allocation function means maximizing the bandwidth resources allocated to a service so that that type of service can be successfully allocated.

[0082] Common business resource allocation function:

[0083] The ordinary service resource allocation function aims to maximize the successful access time of this type of service, so that this type of service can be successfully allocated.

[0084] When network resources are limited, there is a game-theoretic relationship between different objective functions.

[0085] Step 4b: Set constraints for the multi-objective network model:

[0086] business At any given time, bandwidth resources can only be allocated within one time window for a specific inter-satellite link of a specific satellite; bandwidth resources for inter-satellite links cannot be allocated simultaneously across multiple time windows.

[0087] business Resource allocation must be performed within the visible time window of the link.

[0088] Prioritize satisfying the objective function of highly reliable deterministic services, followed by the objective function of high-bandwidth services, and lastly the objective function of ordinary services.

[0089] Step 5: Propose a resource normalization method to convert the multi-objective function of the business into the corresponding fitness function.

[0090] Step 5a: Calculate the total satellite bandwidth resources contained in the network. First, calculate the total inter-satellite link bandwidth resources for each satellite, then sum the bandwidth resources of all satellites to obtain... =376304Mbps.

[0091] Step 5b, Calculate the business Request bandwidth resources Numerical and The ratio, as a business The requested normalized bandwidth resource value. (Based on service) For example, its application for bandwidth resources =310Mbps, its normalized bandwidth resource is the ratio of the requested bandwidth resource to the sum of bandwidth resources, which is calculated to be 8.238×10 -4 .

[0092] Step 5c: Calculate the total business weight of the application business. Summation yields =593.

[0093] Step 5d, Calculation Business Business weight and The ratio, as a business Normalized business weights. Based on business For example, its application business priority is 8, business weight is 32, and normalized business weight is 5.396 × 10. -2 .

[0094] Step 5e: Calculate the total available window time for all satellites in the network. First, calculate the sum of the available time windows for different inter-satellite links for each satellite, then sum the available time windows for all satellites to obtain the result. =363668 (s).

[0095] Step 5f, Calculate the business Application Service Duration and The ratio, as a business The normalized duration of the application. (Based on business...) For example, the duration of its application process =430 (s), and its normalized service duration is the ratio of the requested service duration to the sum of the available satellite time windows, which is calculated to be 1.182 × 10. -3 .

[0096] Step 5g: The multi-objective fitness function of the low-orbit carrier network is obtained as follows:

[0097] Highly reliable deterministic business fitness function:

[0098] High-bandwidth service fitness function:

[0099] Typical business fitness function:

[0100] Step 5h: Set the weight coefficients of the multi-objective fitness function to obtain the fitness function of the low-Earth orbit network:

[0101]

[0102] in , , The fitness function adjustment coefficient can be adjusted according to the weights of the multi-objective function. In this embodiment, it is set as follows: =0.45, =0.3, =0.2.

[0103] Step 6: Using the improved NSGA-II genetic algorithm, perform multi-objective optimization based on the fitness function from Step 5 to obtain the Parto optimal solution. See the flowchart below. Figure 2 .

[0104] Step 6a: Set the population size to 60, randomly generate an initial population, and adapt the business parameter information to the population data format required by the algorithm.

[0105] Step 6b: Calculate the fitness value of each business chromosome in the population based on the fitness function.

[0106] Step 6c: Update satellite available bandwidth resources and time window information.

[0107] Step 6d: Perform non-dominated sorting, calculate the crowding distance of the sorted population, and perform adaptive crossover and mutation operations.

[0108] Step 6e: Generate the next generation population.

[0109] Step 6f: Determine if the current generation is the maximum generation. If not, proceed to step 6b; if so, proceed to step 6g.

[0110] Step 6g: Obtain the Pareto optimal solution set under the current number of genetic iterations and obtain the current resource allocation status.

[0111] Step 7: Based on the resource allocation optimization results obtained by the algorithm, allocate the corresponding satellite bandwidth resources to the applied services.

[0112] Simulations yielded the values ​​of unsuccessfully allocated service weights in the network as the algorithm iterated. Figure 5 ), total bandwidth of services not successfully allocated ( Figure 6 ) and unsuccessfully allocated business time ( Figure 7The trend chart shows that after 100 simulations, the algorithm has converged. Because a normalization method is used here, the percentage of unsuccessfully allocated service weights is 0.13, the percentage of unsuccessfully allocated service bandwidth is 0.311, and the percentage of unsuccessfully allocated service time is 0.12. Note: The denominator used in this algorithm's normalization is the sum of available resources for all satellites (not the sum of requested service resources), while the numerator is the resource request from the access satellite. Even if all services are accessed, the successfully allocated service weights, bandwidth, and service time cannot reach 100%, therefore the probability of failure cannot be 0.

[0113] As can be seen, the method proposed in this invention can specifically meet the QoS requirements of different services.

[0114] In summary, the present invention provides a method for multi-target resource allocation in low-Earth orbit satellite networks with multi-service QoS.

[0115] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope defined in the claims of the present invention.

Claims

1. A method for multi-target resource allocation in low-Earth orbit satellite networks oriented towards multi-service QoS, characterized in that, include: The satellite control center obtains initial status information of the satellite bearer network, including satellite link availability time, total satellite bandwidth resources, and available bandwidth resources. Based on the latency, bandwidth, and QoS requirements of key services in the low-Earth orbit bearer network, and combined with service priorities, corresponding key services are extracted, and key service weights are generated. Based on the spatial area corresponding to the low-Earth orbit satellite carrier network services, select satellites to carry the services, reduce the number of satellites that need to be allocated resources, and establish a multi-objective constrained network model for the low-Earth orbit satellite carrier network. To meet the QoS requirements of different service levels, a multi-objective optimization function is generated for the aforementioned low-Earth orbit satellite bearer network model. By using resource normalization methods, the multi-objective function of the business is converted into a corresponding fitness function; The optimal solution is obtained by solving a multi-objective optimization problem oriented towards the fitness function. Based on the resource allocation optimization results obtained by the algorithm, the corresponding satellite bandwidth resources will be allocated to the applied services; If the business requests a delay Less than the delay threshold The applied latency jitter Less than the delay jitter threshold The requested packet loss rate Less than the packet loss rate threshold If so, the service is classified as a highly reliable deterministic service and denoted as... If the service requests bandwidth Greater than the bandwidth threshold If so, the service is classified as a high-bandwidth service and denoted as... ; All other business is classified as ordinary business and is denoted as ; It is a collection of low-Earth orbit satellite bearer network services, among which For a set of highly reliable deterministic services, For high-bandwidth service collection, It is a collection of ordinary business processes, and ; The multi-objective optimization function is as follows: High-reliability deterministic business resource allocation optimization function: ; :business The converted weights; : This is the business sequence number, with a value range from 1 to... ; Number of low-Earth orbit satellite network services; High-bandwidth service optimized resource allocation function: ; :business Requested bandwidth resource values; Common business resource allocation function: ; :business Request communication time; : The business successfully allocated a resource identifier, if the business If bandwidth resources are successfully allocated, then ,otherwise ; The multi-objective fitness function of the low-orbit carrier network is as follows: Highly reliable deterministic business fitness function: , The sum of the business weights for the applied business; High-bandwidth service fitness function: , This represents the total satellite bandwidth resources contained in the network. Typical business fitness function: , This represents the total available window time for all satellites included in the network. Constraints of the multi-objective network model: A service can only allocate bandwidth resources within one time window of a specific inter-satellite link on a specific satellite at any given time; it cannot allocate inter-satellite link bandwidth resources simultaneously across multiple time windows. Businesses must allocate resources within the visible time window of the link; Prioritize satisfying the objective function of highly reliable deterministic services, followed by the objective function of high-bandwidth services, and lastly the objective function of ordinary services.

2. The method according to claim 1, characterized in that, The final weight of each business is recorded as: The highest priority corresponding to each type of priority matrix is ​​denoted as The weighting coefficients for each type of business are set as follows: The value obtained by rounding down from 2 is represented as the business weight. ; :business Application priority; The weighting coefficients for each type of business are as follows: , and ,Depend on The value obtained by rounding down by 2; : This is the business type number, with a value range from 1 to 3.

3. The method according to claim 1, characterized in that, Low-orbit bearing network fitness function ,in , , This is the adjustment coefficient of the fitness function. , , For multi-objective fitness functions for different business operations.

4. The method according to claim 1, characterized in that, The multi-objective optimization solution oriented towards the fitness function, to obtain the optimal solution, includes: An improved NSGA-II genetic algorithm is used to perform multi-objective optimization based on the fitness function, and the Parto optimal solution is obtained.

5. The method according to claim 4, characterized in that, The method includes: Set the population size, randomly generate an initial population, and adapt the business parameter information to the population data format required by the algorithm. The fitness value of each operational chromosome in the population is calculated based on the fitness function; Update satellite available bandwidth resources and time window information; Perform non-dominated sorting, calculate the crowding distance of the sorted population, and perform adaptive crossover and mutation operations; Generate the next generation of population; Determine if the current generation is the maximum generation. If not, recalculate the fitness value of each business chromosome. If it is, obtain the Pareto optimal solution set under the current generation and obtain the current resource allocation status.

Images