Satellite Network Resource Allocation via Service Ordering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing resource allocation methods for large-scale satellite networks (LSNs) fail to effectively consider the impact of service order on overall service performance and struggle with scalability as network scales expand.
Innovation Solution
A service-level communication and computing collaborative resource allocation method and device that utilizes a reinforcement learning framework, incorporating a service order decision submodule and a pointer network to determine the service order and select resource allocation strategies independently of network scale.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing resource allocation methods are used for large-scale satellite networks, then the basic service allocation can be performed, but the service performance is degraded due to not considering service order impact and the scalability deteriorates as network scale expands
Solution Approach 1:
The patent segments the resource allocation problem into two distinct modules: service order determination module and resource allocation strategy selection module. This segmentation allows each module to be optimized independently, improving service performance through proper service ordering while maintaining scalability through independent strategy selection that doesn't depend on network scale
Solution Approach 2:
The patent introduces service order as an intermediary element between service requests and resource allocation strategies. This intermediary optimizes the sequence of service processing to improve overall service performance, while the resource allocation strategies remain scalable and independent of network size
2Device complexity
If traditional resource allocation methods are used, then the implementation is relatively simple, but the solution complexity increases and scalability deteriorates as network service demands, nodes, and links expand
Solution Approach 1:
By segmenting the allocation method into service order determination and strategy selection, the patent reduces overall solution complexity. Each segment can be processed independently, avoiding the combinatorial explosion that would occur in traditional methods as network scale expands
Solution Approach 2:
The patent performs preliminary service order determination before resource allocation strategy selection. This preliminary action optimizes the service sequence in advance, simplifying subsequent allocation decisions and reducing overall computational complexity regardless of network scale
Data Source
AI summary
A service-level communication and computing collaborative resource allocation device and method for a large-scale satellite network are provided, the device includes: a network information management function module including a service request collection unit configured to collect service requests arriving at each satellite node in the network, and a network resource collection unit configured to collect available resource status information for each satellite node; a service preprocessing function module configured to preprocess service requests; a service feature extraction function module configured to perform feature extraction for preprocessed service requests to generate service demand features; a service order decision submodule configured to determine a service order of service requests; a service strategy decision submodule configured to output a service strategy of the service request. The device and method can effectively ensure that the large-scale satellite network provides high-performance services for service with service demands of communication and computing collaborative resources.