Meshed Wireless Network Packet Scheduling via Cost Function Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing meshed wireless network scheduling schemes face challenges in achieving fair and efficient link prioritization, particularly in multi-hop scenarios, where randomization is needed to ensure fairness over time, leading to suboptimal resource utilization and latency.
Innovation Solution
A method for forwarding data packets in meshed wireless networks that determines a cost function for each packet to select optimal next hop nodes and time slots, maximizing medium utilization and minimizing latency by assigning packets to selected time slots and nodes, ensuring joint optimization and preventing packet starvation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If links are assigned priorities in a distributed scheduling scheme, then higher priority links can be scheduled first, but randomization is needed to achieve fairness over time, leading to suboptimal resource utilization
Solution Approach 1:
The patent pre-calculates and stores cost functions for all possible next-hop nodes and time slots before actual packet transmission. This preliminary computation includes the number of time slots required to reach the destination and the waiting time before transmission can occur. By having these cost values ready in advance, the scheduling decision can be made immediately without randomization or complex real-time prioritization, thus reducing latency while maintaining optimal resource utilization.
Solution Approach 2:
Each transmitting node independently determines the cost function for its own packets and makes autonomous scheduling decisions without requiring network-wide randomization or centralized coordination. The node uses the pre-calculated cost values to select the optimal next-hop and time slot combination, achieving fairness and efficiency through local optimization rather than global randomization.
2Ease of operation
If packets are scheduled individually without joint optimization, then scheduling is simpler, but packet starvation can occur for particular destinations
Solution Approach 1:
The patent merges the scheduling decisions for multiple packets into a single joint optimization process. Instead of scheduling packets individually, the system considers all queued packets simultaneously and assigns them to time slots and next-hop nodes together, using the sum of cost functions as the optimization criterion. This ensures that no single destination is starved, as the joint optimization balances resource allocation across all destinations while maintaining manageable complexity through the use of pre-calculated cost values.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
The invention pertains to a method for forwarding a data packet in a meshed wireless communication network comprising a plurality of nodes that are only permitted to transmit in accordance with a predetermined transmission schedule such that different sets of transmitter-receiver pairs are active in different time slots, the method comprising at a transmitting node having a set of packets queued for transmission: selecting (1150) from the nodes that can be reached during the active time slot according to said predetermined transmission schedule, a next hop node according to a ranking; selecting (1160) from said set of packets, a packet having a destination for which said selected next hop meets a predetermined destination criterion; and transmitting (1170) said selected packet to said selected next hop node.